Add files using upload-large-folder tool

.github/workflows/clean-host-contract.yml

@@ -1,35 +1,35 @@
-name: Clean Host Contract
-
-on:
-  workflow_dispatch:
-  schedule:
-    - cron: "17 3 * * 1"
-
-env:
-  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
-
-permissions:
-  contents: read
-
-jobs:
-  clean-host-contract:
-    name: Clean wheel install and A-Z contract
-    runs-on: ubuntu-latest
-    timeout-minutes: 25
-
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-
-      - name: Set up Python 3.12
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.12"
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Upgrade pip
-        run: python -m pip install --upgrade pip
-
-      - name: Run clean-host contract
-        run: python scripts/clean_host_contract.py --fast
+name: Clean Host Contract
+
+on:
+  workflow_dispatch:
+  schedule:
+    - cron: "17 3 * * 1"
+
+env:
+  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
+
+permissions:
+  contents: read
+
+jobs:
+  clean-host-contract:
+    name: Clean wheel install and A-Z contract
+    runs-on: ubuntu-latest
+    timeout-minutes: 25
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+
+      - name: Set up Python 3.12
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.12"
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Upgrade pip
+        run: python -m pip install --upgrade pip
+
+      - name: Run clean-host contract
+        run: python scripts/clean_host_contract.py --fast

.github/workflows/docs.yml

@@ -1,73 +1,73 @@
-name: Deploy docs to GitHub Pages
-
-# Build the MkDocs Material site and publish to GitHub Pages on every push
-# to main. Also deployable on demand via the workflow_dispatch trigger.
-
-on:
-  push:
-    branches:
-      - main
-    paths:
-      - "docs/**"
-      - "mkdocs.yml"
-      - "requirements-docs.txt"
-      - ".github/workflows/docs.yml"
-  workflow_dispatch:
-
-env:
-  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
-
-permissions:
-  contents: read
-
-# Allow only one concurrent deploy; cancel in-progress runs on a new push.
-concurrency:
-  group: "pages"
-  cancel-in-progress: true
-
-jobs:
-  build:
-    name: Build site
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          fetch-depth: 0   # Required by mkdocs git-revision plugins if added.
-
-      - name: Set up Python
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.11"
-          cache: "pip"
-          cache-dependency-path: requirements-docs.txt
-
-      - name: Install docs dependencies
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install -r requirements-docs.txt
-
-      - name: Build site (strict)
-        run: |
-          python -m mkdocs build --strict
-
-      - name: Upload Pages artifact
-        uses: actions/upload-pages-artifact@v5
-        with:
-          path: site
-
-  deploy:
-    name: Deploy to GitHub Pages
-    needs: build
-    runs-on: ubuntu-latest
-    permissions:
-      contents: read
-      pages: write
-      id-token: write
-    environment:
-      name: github-pages
-      url: ${{ steps.deployment.outputs.page_url }}
-    steps:
-      - name: Deploy
-        id: deployment
-        uses: actions/deploy-pages@v5
+name: Deploy docs to GitHub Pages
+
+# Build the MkDocs Material site and publish to GitHub Pages on every push
+# to main. Also deployable on demand via the workflow_dispatch trigger.
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - "docs/**"
+      - "mkdocs.yml"
+      - "requirements-docs.txt"
+      - ".github/workflows/docs.yml"
+  workflow_dispatch:
+
+env:
+  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
+
+permissions:
+  contents: read
+
+# Allow only one concurrent deploy; cancel in-progress runs on a new push.
+concurrency:
+  group: "pages"
+  cancel-in-progress: true
+
+jobs:
+  build:
+    name: Build site
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          fetch-depth: 0   # Required by mkdocs git-revision plugins if added.
+
+      - name: Set up Python
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+          cache: "pip"
+          cache-dependency-path: requirements-docs.txt
+
+      - name: Install docs dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install -r requirements-docs.txt
+
+      - name: Build site (strict)
+        run: |
+          python -m mkdocs build --strict
+
+      - name: Upload Pages artifact
+        uses: actions/upload-pages-artifact@v5
+        with:
+          path: site
+
+  deploy:
+    name: Deploy to GitHub Pages
+    needs: build
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      pages: write
+      id-token: write
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+    steps:
+      - name: Deploy
+        id: deployment
+        uses: actions/deploy-pages@v5

.github/workflows/huggingface-sync.yml

@@ -0,0 +1,64 @@
+name: Sync Hugging Face
+
+on:
+  push:
+    branches:
+      - main
+  workflow_dispatch:
+
+env:
+  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
+  HF_TOKEN: ${{ secrets.HF_TOKEN }}
+
+permissions:
+  contents: read
+
+concurrency:
+  group: "huggingface-sync-main"
+  cancel-in-progress: true
+
+jobs:
+  sync:
+    name: Sync dataset
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Skip when HF_TOKEN is not configured
+        if: ${{ env.HF_TOKEN == '' }}
+        run: |
+          echo "::notice title=Hugging Face sync skipped::Set the HF_TOKEN repository secret to publish Stevesolun/ctx automatically."
+
+      - name: Hydrate required graph artifacts from release assets
+        if: ${{ env.HF_TOKEN != '' }}
+        env:
+          GH_TOKEN: ${{ github.token }}
+        run: |
+          latest_tag="$(gh release view --repo "$GITHUB_REPOSITORY" --json tagName --jq .tagName)"
+          echo "Hydrating graph artifacts from release ${latest_tag}"
+          rm -f graph/wiki-graph.tar.gz graph/wiki-graph-runtime.tar.gz graph/skills-sh-catalog.json.gz
+          gh release download "$latest_tag" --repo "$GITHUB_REPOSITORY" --dir graph --pattern wiki-graph.tar.gz
+          gh release download "$latest_tag" --repo "$GITHUB_REPOSITORY" --dir graph --pattern wiki-graph-runtime.tar.gz
+          gh release download "$latest_tag" --repo "$GITHUB_REPOSITORY" --dir graph --pattern skills-sh-catalog.json.gz
+
+      - name: Set up Python
+        if: ${{ env.HF_TOKEN != '' }}
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+          cache: "pip"
+          cache-dependency-path: pyproject.toml
+
+      - name: Install sync dependencies
+        if: ${{ env.HF_TOKEN != '' }}
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install -e ".[dev]" huggingface_hub
+
+      - name: Sync Hugging Face dataset
+        if: ${{ env.HF_TOKEN != '' }}
+        run: |
+          python scripts/sync_huggingface.py --repo . --repo-id Stevesolun/ctx --repo-type dataset

.github/workflows/publish.yml

@@ -1,475 +1,481 @@
-name: Publish to PyPI
-
-# Publish claude-ctx to PyPI on every tag that looks like a version
-# (v0.5.0, v0.5.0-rc1, v1.0.0, …). Uses PyPI Trusted Publishing —
-# no API token needed once the pending publisher is configured at
-# https://pypi.org/manage/account/publishing/.
-#
-# Also deployable on demand to TestPyPI via workflow_dispatch. Production
-# PyPI publishes must use a version tag push so graph release assets are
-# uploaded before the wheel is published.
-
-on:
-  push:
-    tags:
-      - "v*"
-  workflow_dispatch:
-    inputs:
-      repository:
-        description: "PyPI repository (pypi or testpypi)"
-        required: true
-        default: "pypi"
-        type: choice
-        options:
-          - pypi
-          - testpypi
-
-env:
-  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
-
-permissions:
-  contents: read
-
-jobs:
-  build:
-    name: Build sdist + wheel
-    runs-on: ubuntu-latest
-    permissions:
-      contents: read
-    outputs:
-      graph_assets_available: ${{ steps.resolve_graph_lfs.outcome == 'success' }}
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.12"
-
-      - name: Install release tooling
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install ".[dev]" build packaging twine
-
-      - name: Validate release target
-        run: |
-          python - <<'PY'
-          import os
-          import tomllib
-          from packaging.version import Version
-
-          event = os.environ["GITHUB_EVENT_NAME"]
-          target_repository = os.environ.get("INPUT_REPOSITORY", "pypi")
-          ref_type = os.environ.get("GITHUB_REF_TYPE", "")
-          ref_name = os.environ["GITHUB_REF_NAME"]
-
-          if event == "workflow_dispatch" and target_repository == "pypi":
-              raise SystemExit(
-                  "Manual PyPI publish is disabled; push a version tag so "
-                  "graph release assets are uploaded before publishing"
-              )
-          if event == "workflow_dispatch" and target_repository == "testpypi":
-              print(f"manual TestPyPI publish allowed from {ref_type}:{ref_name}")
-              raise SystemExit(0)
-
-          if ref_type != "tag":
-              raise SystemExit(
-                  "PyPI release must run from a version tag; "
-                  f"got {ref_type}:{ref_name}"
-              )
-          tag = ref_name
-          if not tag.startswith("v"):
-              raise SystemExit(f"release tag must start with v: {tag}")
-
-          with open("pyproject.toml", "rb") as fh:
-              package_version = tomllib.load(fh)["project"]["version"]
-
-          tag_version = str(Version(tag[1:]))
-          normalized_package_version = str(Version(package_version))
-          if tag_version != normalized_package_version:
-              raise SystemExit(
-                  f"tag {tag!r} does not match pyproject version {package_version!r}"
-              )
-          print(f"release version {package_version} matches tag {tag}")
-          PY
-        env:
-          INPUT_REPOSITORY: ${{ github.event.inputs.repository || 'pypi' }}
-
-      - name: Reject already published PyPI version
-        run: |
-          python - <<'PY'
-          import os
-          import tomllib
-          import urllib.error
-          import urllib.request
-
-          target_repository = os.environ.get("INPUT_REPOSITORY", "pypi")
-          if target_repository != "pypi":
-              print(f"skipping PyPI reuse check for {target_repository}")
-              raise SystemExit(0)
-
-          with open("pyproject.toml", "rb") as fh:
-              project = tomllib.load(fh)["project"]
-
-          name = project["name"]
-          package_version = project["version"]
-          url = f"https://pypi.org/pypi/{name}/{package_version}/json"
-          try:
-              with urllib.request.urlopen(url, timeout=15):
-                  raise SystemExit(
-                      f"{name} {package_version} already exists on PyPI; "
-                      "bump the version before publishing"
-                  )
-          except urllib.error.HTTPError as exc:
-              if exc.code == 404:
-                  print(f"{name} {package_version} is not present on PyPI")
-                  raise SystemExit(0)
-              raise
-          PY
-        env:
-          INPUT_REPOSITORY: ${{ github.event.inputs.repository || 'pypi' }}
-
-      - name: Resolve release graph LFS artifacts
-        id: resolve_graph_lfs
-        env:
-          GH_TOKEN: ${{ github.token }}
-          TAG_NAME: ${{ github.ref_name }}
-        run: |
-          set -euo pipefail
-          git lfs install --local
-          if git lfs pull --include="graph/wiki-graph.tar.gz,graph/wiki-graph-runtime.tar.gz" --exclude=""; then
-            exit 0
-          fi
-          echo "Git LFS download failed; trying matching prior release asset."
-          python - <<'PY'
-          import hashlib
-          import json
-          import os
-          from pathlib import Path
-          import subprocess
-          import time
-          import urllib.request
-
-          repo = os.environ["GITHUB_REPOSITORY"]
-          current_tag = os.environ.get("TAG_NAME", "")
-          release_asset_wait_seconds = 300
-          release_asset_poll_seconds = 10
-
-          def load_releases() -> list[dict]:
-              return json.loads(subprocess.check_output(
-                  ["gh", "api", f"repos/{repo}/releases?per_page=50"],
-                  text=True,
-              ))
-
-          def hydrate_from_release(path_name: str, hydrated_min_size: int) -> None:
-              graph_tar = Path(path_name)
-              pointer = graph_tar.read_text(encoding="utf-8", errors="replace")
-              expected_oid = ""
-              expected_size = 0
-              for line in pointer.splitlines():
-                  if line.startswith("oid sha256:"):
-                      expected_oid = line.split(":", 1)[1].strip()
-                  elif line.startswith("size "):
-                      expected_size = int(line.split(" ", 1)[1].strip())
-              if not expected_oid:
-                  if graph_tar.stat().st_size > hydrated_min_size:
-                      print(f"{graph_tar} is already hydrated")
-                      return
-                  raise SystemExit(f"{path_name} is neither hydrated nor an LFS pointer")
-
-              deadline = time.monotonic() + release_asset_wait_seconds
-              while True:
-                  candidates = []
-                  for release in load_releases():
-                      tag_name = str(release.get("tag_name") or "")
-                      is_graph_cache = tag_name.startswith("graph-artifacts-")
-                      if release.get("draft") or (
-                          release.get("prerelease") and not is_graph_cache
-                      ):
-                          continue
-                      if tag_name == current_tag:
-                          continue
-                      for asset in release.get("assets", []):
-                          if asset.get("name") != graph_tar.name:
-                              continue
-                          digest = str(asset.get("digest") or "")
-                          size = int(asset.get("size") or 0)
-                          if size != expected_size:
-                              continue
-                          if digest and digest != f"sha256:{expected_oid}":
-                              continue
-                          candidates.append((tag_name, asset))
-
-                  if candidates:
-                      break
-                  if time.monotonic() >= deadline:
-                      raise SystemExit(
-                          f"No previous release asset matches {path_name} "
-                          f"sha256:{expected_oid} size:{expected_size}"
-                      )
-                  print(
-                      f"Waiting for matching release asset {graph_tar.name} "
-                      f"sha256:{expected_oid} size:{expected_size}"
-                  )
-                  time.sleep(release_asset_poll_seconds)
-
-              source_tag, asset = candidates[0]
-              tmp = graph_tar.with_name(f"{graph_tar.name}.download")
-              sha = hashlib.sha256()
-              total = 0
-              with urllib.request.urlopen(asset["browser_download_url"], timeout=300) as resp:  # noqa: S310
-                  with tmp.open("wb") as fh:
-                      while True:
-                          chunk = resp.read(1024 * 1024)
-                          if not chunk:
-                              break
-                          sha.update(chunk)
-                          total += len(chunk)
-                          fh.write(chunk)
-              actual_oid = sha.hexdigest()
-              if actual_oid != expected_oid or total != expected_size:
-                  tmp.unlink(missing_ok=True)
-                  raise SystemExit(
-                      f"Downloaded {path_name} does not match LFS pointer: "
-                      f"sha256:{actual_oid} size:{total}"
-                  )
-              tmp.replace(graph_tar)
-              print(
-                  f"Hydrated {path_name} from {source_tag} release asset "
-                  f"sha256:{actual_oid} size:{total}"
-              )
-
-          hydrate_from_release("graph/wiki-graph.tar.gz", 100_000_000)
-          hydrate_from_release("graph/wiki-graph-runtime.tar.gz", 10_000_000)
-          PY
-
-      - name: Validate release graph artifacts
-        if: steps.resolve_graph_lfs.outcome == 'success'
-        run: |
-          python src/validate_graph_artifacts.py \
-            --graph-dir graph \
-            --deep \
-            --min-nodes 100000 \
-            --min-edges 2000000 \
-            --min-skills-sh-nodes 89000 \
-            --min-semantic-edges 1000000 \
-            --expected-nodes 102928 \
-            --expected-edges 2913960 \
-            --expected-semantic-edges 1683193 \
-            --expected-harness-nodes 207 \
-            --expected-skills-sh-nodes 89471 \
-            --expected-skills-sh-catalog-entries 89465 \
-            --expected-skills-sh-converted 89465 \
-            --expected-skill-pages 91464 \
-            --expected-agent-pages 467 \
-            --expected-mcp-pages 10790 \
-            --expected-harness-pages 207 \
-            --line-threshold 180 \
-            --max-stage-lines 40
-
-      - name: Validate README and docs stats
-        if: steps.resolve_graph_lfs.outcome == 'success'
-        run: python src/update_repo_stats.py --check
-
-      - name: Static gates
-        run: |
-          python -m ruff check src hooks scripts
-          python -m mypy src
-
-      - name: Clean-host contract
-        run: python scripts/clean_host_contract.py --fast
-
-      - name: Release canary tests
-        run: |
-          python -m pytest -q --no-cov \
-            src/tests/test_ci_classifier.py \
-            src/tests/test_package_scaffold.py \
-            src/tests/test_clean_host_contract.py \
-            src/tests/test_alive_loop_e2e.py
-
-      - name: Build distributions
-        run: python -m build
-
-      - name: Check distributions
-        run: python -m twine check dist/*
-
-      - name: Check distribution contents
-        run: |
-          python - <<'PY'
-          import tarfile
-          import zipfile
-          from pathlib import Path
-
-          wheel = next(Path("dist").glob("*.whl"))
-          sdist = next(Path("dist").glob("*.tar.gz"))
-
-          with zipfile.ZipFile(wheel) as zf:
-              wheel_names = set(zf.namelist())
-          required = {"ctx/config.json", "ctx/skill-registry.json"}
-          missing = sorted(required - wheel_names)
-          if missing:
-              raise SystemExit(f"wheel missing packaged defaults: {missing}")
-
-          with tarfile.open(sdist, "r:gz") as tf:
-              names = set(tf.getnames())
-          forbidden = ("/src/tests/", "/.claude/", "/.a5c/")
-          leaked = sorted(name for name in names if any(part in name for part in forbidden))
-          if leaked:
-              raise SystemExit("sdist contains local/test artifacts:\n" + "\n".join(leaked[:20]))
-          print(f"checked distribution contents: {wheel.name}, {sdist.name}")
-          PY
-
-      - name: Smoke install wheel
-        run: |
-          python -m venv .venv-smoke
-          . .venv-smoke/bin/activate
-          python -m pip install --upgrade pip
-          wheel="$(python - <<'PY'
-          from pathlib import Path
-          print(next(Path("dist").glob("*.whl")))
-          PY
-          )"
-          python -m pip install "$wheel"
-          python -m pip check
-          python - <<'PY'
-          import tomllib
-          from importlib.metadata import entry_points, version
-
-          import ctx
-          import ctx_config
-
-          with open("pyproject.toml", "rb") as fh:
-              expected_scripts = set(tomllib.load(fh)["project"]["scripts"])
-
-          dist_version = version("claude-ctx")
-          if ctx.__version__ != dist_version:
-              raise SystemExit(
-                  f"ctx.__version__={ctx.__version__!r} != metadata {dist_version!r}"
-              )
-          if ctx_config.cfg.recommendation_top_k != 5:
-              raise SystemExit("packaged default config was not loaded")
-          scripts = {
-              ep.name: ep for ep in entry_points(group="console_scripts")
-              if ep.name == "ctx" or ep.name.startswith("ctx-")
-          }
-          missing = sorted(expected_scripts - set(scripts))
-          extra = sorted(set(scripts) - expected_scripts)
-          if missing or extra:
-              raise SystemExit(
-                  "wheel console-script surface mismatch\n"
-                  f"missing: {missing}\n"
-                  f"extra: {extra}"
-              )
-          failures = []
-          for ep in scripts.values():
-              try:
-                  ep.load()
-              except Exception as exc:
-                  failures.append(f"{ep.name}: {exc!r}")
-          if failures:
-              raise SystemExit("console script load failures:\n" + "\n".join(failures))
-          unsafe_help = {"ctx-mcp-server"}
-          safe_help = sorted(expected_scripts - unsafe_help)
-          with open("ctx-console-help.txt", "w", encoding="utf-8", newline="\n") as fh:
-              fh.write("\n".join(safe_help) + "\n")
-          print(
-              f"loaded {len(scripts)} ctx console scripts from wheel "
-              f"{dist_version}; help-smoke={len(safe_help)}"
-          )
-          PY
-          while IFS= read -r cmd; do
-            cmd="${cmd%$'\r'}"
-            [[ -z "$cmd" ]] && continue
-            echo "help smoke: $cmd"
-            "$cmd" --help >/dev/null
-          done < ctx-console-help.txt
-          python -m pip install "${wheel}[harness]"
-          python - <<'PY'
-          import litellm
-
-          print(f"harness extra import ok: litellm {getattr(litellm, '__version__', 'unknown')}")
-          PY
-
-      - name: Upload dist artifact
-        uses: actions/upload-artifact@v7
-        with:
-          name: dist
-          path: dist/
-
-      - name: Upload graph artifact bundle
-        if: steps.resolve_graph_lfs.outcome == 'success'
-        uses: actions/upload-artifact@v7
-        with:
-          name: graph-release-assets
-          path: |
-            graph/wiki-graph.tar.gz
-            graph/wiki-graph-runtime.tar.gz
-            graph/skills-sh-catalog.json.gz
-            graph/communities.json
-            graph/entity-overlays.jsonl
-
-  release-assets:
-    name: Upload graph release assets
-    needs: build
-    if: github.event_name == 'push' && needs.build.outputs.graph_assets_available == 'true'
-    runs-on: ubuntu-latest
-    permissions:
-      contents: write
-    steps:
-      - name: Download graph artifact bundle
-        uses: actions/download-artifact@v7
-        with:
-          name: graph-release-assets
-          path: graph-release-assets
-
-      - name: Upload graph assets to GitHub release
-        env:
-          GH_TOKEN: ${{ github.token }}
-          TAG_NAME: ${{ github.ref_name }}
-        run: |
-          gh release view "$TAG_NAME" --repo "$GITHUB_REPOSITORY" >/dev/null 2>&1 \
-            || gh release create "$TAG_NAME" \
-              --repo "$GITHUB_REPOSITORY" \
-              --title "$TAG_NAME" \
-              --notes "ctx $TAG_NAME"
-          gh release upload "$TAG_NAME" \
-            --repo "$GITHUB_REPOSITORY" \
-            graph-release-assets/wiki-graph.tar.gz \
-            graph-release-assets/wiki-graph-runtime.tar.gz \
-            graph-release-assets/skills-sh-catalog.json.gz \
-            graph-release-assets/communities.json \
-            graph-release-assets/entity-overlays.jsonl \
-            --clobber
-
-  publish:
-    name: Publish to PyPI
-    needs:
-      - build
-      - release-assets
-    if: ${{ always() && needs.build.result == 'success' && (needs.release-assets.result == 'success' || (github.event_name == 'workflow_dispatch' && github.event.inputs.repository == 'testpypi')) }}
-    runs-on: ubuntu-latest
-    permissions:
-      contents: read
-      id-token: write   # required for Trusted Publishing
-    environment:
-      name: pypi
-      url: https://pypi.org/project/claude-ctx/
-    steps:
-      - name: Download dist artifact
-        uses: actions/download-artifact@v7
-        with:
-          name: dist
-          path: dist/
-
-      - name: Publish to PyPI
-        if: github.event_name == 'push'
-        uses: pypa/gh-action-pypi-publish@release/v1
-
-      - name: Publish to TestPyPI
-        if: github.event_name == 'workflow_dispatch' && github.event.inputs.repository == 'testpypi'
-        uses: pypa/gh-action-pypi-publish@release/v1
-        with:
-          repository-url: https://test.pypi.org/legacy/
+name: Publish to PyPI
+
+# Publish claude-ctx to PyPI on every tag that looks like a version
+# (v0.5.0, v0.5.0-rc1, v1.0.0, …). Uses PyPI Trusted Publishing —
+# no API token needed once the pending publisher is configured at
+# https://pypi.org/manage/account/publishing/.
+#
+# Also deployable on demand to TestPyPI via workflow_dispatch. Production
+# PyPI publishes must use a version tag push so graph release assets are
+# uploaded before the wheel is published.
+
+on:
+  push:
+    tags:
+      - "v*"
+  workflow_dispatch:
+    inputs:
+      repository:
+        description: "PyPI repository (pypi or testpypi)"
+        required: true
+        default: "pypi"
+        type: choice
+        options:
+          - pypi
+          - testpypi
+
+env:
+  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
+
+permissions:
+  contents: read
+
+jobs:
+  build:
+    name: Build sdist + wheel
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+    outputs:
+      graph_assets_available: ${{ steps.resolve_graph_assets.outcome == 'success' }}
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.12"
+
+      - name: Install release tooling
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install ".[dev]" build packaging twine
+
+      - name: Validate release target
+        run: |
+          python - <<'PY'
+          import os
+          import tomllib
+          from packaging.version import Version
+
+          event = os.environ["GITHUB_EVENT_NAME"]
+          target_repository = os.environ.get("INPUT_REPOSITORY", "pypi")
+          ref_type = os.environ.get("GITHUB_REF_TYPE", "")
+          ref_name = os.environ["GITHUB_REF_NAME"]
+
+          if event == "workflow_dispatch" and target_repository == "pypi":
+              raise SystemExit(
+                  "Manual PyPI publish is disabled; push a version tag so "
+                  "graph release assets are uploaded before publishing"
+              )
+          if event == "workflow_dispatch" and target_repository == "testpypi":
+              print(f"manual TestPyPI publish allowed from {ref_type}:{ref_name}")
+              raise SystemExit(0)
+
+          if ref_type != "tag":
+              raise SystemExit(
+                  "PyPI release must run from a version tag; "
+                  f"got {ref_type}:{ref_name}"
+              )
+          tag = ref_name
+          if not tag.startswith("v"):
+              raise SystemExit(f"release tag must start with v: {tag}")
+
+          with open("pyproject.toml", "rb") as fh:
+              package_version = tomllib.load(fh)["project"]["version"]
+
+          tag_version = str(Version(tag[1:]))
+          normalized_package_version = str(Version(package_version))
+          if tag_version != normalized_package_version:
+              raise SystemExit(
+                  f"tag {tag!r} does not match pyproject version {package_version!r}"
+              )
+          print(f"release version {package_version} matches tag {tag}")
+          PY
+        env:
+          INPUT_REPOSITORY: ${{ github.event.inputs.repository || 'pypi' }}
+
+      - name: Reject already published PyPI version
+        run: |
+          python - <<'PY'
+          import os
+          import tomllib
+          import urllib.error
+          import urllib.request
+
+          target_repository = os.environ.get("INPUT_REPOSITORY", "pypi")
+          if target_repository != "pypi":
+              print(f"skipping PyPI reuse check for {target_repository}")
+              raise SystemExit(0)
+
+          with open("pyproject.toml", "rb") as fh:
+              project = tomllib.load(fh)["project"]
+
+          name = project["name"]
+          package_version = project["version"]
+          url = f"https://pypi.org/pypi/{name}/{package_version}/json"
+          try:
+              with urllib.request.urlopen(url, timeout=15):
+                  raise SystemExit(
+                      f"{name} {package_version} already exists on PyPI; "
+                      "bump the version before publishing"
+                  )
+          except urllib.error.HTTPError as exc:
+              if exc.code == 404:
+                  print(f"{name} {package_version} is not present on PyPI")
+                  raise SystemExit(0)
+              raise
+          PY
+        env:
+          INPUT_REPOSITORY: ${{ github.event.inputs.repository || 'pypi' }}
+
+      - name: Resolve release graph artifacts from release assets
+        id: resolve_graph_assets
+        env:
+          GH_TOKEN: ${{ github.token }}
+          TAG_NAME: ${{ github.ref_name }}
+        run: |
+          set -euo pipefail
+          echo "Resolving graph artifacts from matching release assets to avoid Git LFS bandwidth."
+          python - <<'PY'
+          import hashlib
+          import json
+          import os
+          from pathlib import Path
+          import subprocess
+          import time
+          import urllib.request
+
+          repo = os.environ["GITHUB_REPOSITORY"]
+          current_tag = os.environ.get("TAG_NAME", "")
+          release_asset_wait_seconds = 300
+          release_asset_poll_seconds = 10
+          expected_graph_assets = {
+              "graph/wiki-graph.tar.gz": {
+                  "sha256": "91b30795e7d200cf31a62a8749969d12658f5f74636d2de06d6b2b24b393c12f",
+                  "size": 457545675,
+              },
+              "graph/wiki-graph-runtime.tar.gz": {
+                  "sha256": "334fb19bace3fd6e4b92087850f17297fb248032957d123f3f1432dfde2e36c0",
+                  "size": 175773376,
+              },
+          }
+
+          def load_releases() -> list[dict]:
+              return json.loads(subprocess.check_output(
+                  ["gh", "api", f"repos/{repo}/releases?per_page=50"],
+                  text=True,
+              ))
+
+          def hydrate_from_release(path_name: str, hydrated_min_size: int) -> None:
+              graph_tar = Path(path_name)
+              fallback = expected_graph_assets[path_name]
+              expected_oid = fallback["sha256"]
+              expected_size = int(fallback["size"])
+              if graph_tar.exists():
+                  pointer = graph_tar.read_text(encoding="utf-8", errors="replace")
+                  for line in pointer.splitlines():
+                      if line.startswith("oid sha256:"):
+                          expected_oid = line.split(":", 1)[1].strip()
+                      elif line.startswith("size "):
+                          expected_size = int(line.split(" ", 1)[1].strip())
+                  if not pointer.startswith("version https://git-lfs.github.com/spec/v1") and graph_tar.stat().st_size > hydrated_min_size:
+                      print(f"{graph_tar} is already hydrated")
+                      return
+
+              deadline = time.monotonic() + release_asset_wait_seconds
+              while True:
+                  candidates = []
+                  for release in load_releases():
+                      tag_name = str(release.get("tag_name") or "")
+                      is_graph_cache = tag_name.startswith("graph-artifacts-")
+                      if release.get("draft") or (
+                          release.get("prerelease") and not is_graph_cache
+                      ):
+                          continue
+                      if tag_name == current_tag:
+                          continue
+                      for asset in release.get("assets", []):
+                          if asset.get("name") != graph_tar.name:
+                              continue
+                          digest = str(asset.get("digest") or "")
+                          size = int(asset.get("size") or 0)
+                          if size != expected_size:
+                              continue
+                          if digest and digest != f"sha256:{expected_oid}":
+                              continue
+                          candidates.append((tag_name, asset))
+
+                  if candidates:
+                      break
+                  if time.monotonic() >= deadline:
+                      raise SystemExit(
+                          f"No previous release asset matches {path_name} "
+                          f"sha256:{expected_oid} size:{expected_size}"
+                      )
+                  print(
+                      f"Waiting for matching release asset {graph_tar.name} "
+                      f"sha256:{expected_oid} size:{expected_size}"
+                  )
+                  time.sleep(release_asset_poll_seconds)
+
+              source_tag, asset = candidates[0]
+              tmp = graph_tar.with_name(f"{graph_tar.name}.download")
+              sha = hashlib.sha256()
+              total = 0
+              with urllib.request.urlopen(asset["browser_download_url"], timeout=300) as resp:  # noqa: S310
+                  with tmp.open("wb") as fh:
+                      while True:
+                          chunk = resp.read(1024 * 1024)
+                          if not chunk:
+                              break
+                          sha.update(chunk)
+                          total += len(chunk)
+                          fh.write(chunk)
+              actual_oid = sha.hexdigest()
+              if actual_oid != expected_oid or total != expected_size:
+                  tmp.unlink(missing_ok=True)
+                  raise SystemExit(
+                      f"Downloaded {path_name} does not match LFS pointer: "
+                      f"sha256:{actual_oid} size:{total}"
+                  )
+              tmp.replace(graph_tar)
+              print(
+                  f"Hydrated {path_name} from {source_tag} release asset "
+                  f"sha256:{actual_oid} size:{total}"
+              )
+
+          hydrate_from_release("graph/wiki-graph.tar.gz", 100_000_000)
+          hydrate_from_release("graph/wiki-graph-runtime.tar.gz", 10_000_000)
+          PY
+
+      - name: Validate release graph artifacts
+        if: steps.resolve_graph_assets.outcome == 'success'
+        run: |
+          python src/validate_graph_artifacts.py \
+            --graph-dir graph \
+            --deep \
+            --min-nodes 100000 \
+            --min-edges 2000000 \
+            --min-skills-sh-nodes 89000 \
+            --min-semantic-edges 1000000 \
+            --expected-nodes 102928 \
+            --expected-edges 2913960 \
+            --expected-semantic-edges 1683193 \
+            --expected-harness-nodes 207 \
+            --expected-skills-sh-nodes 89471 \
+            --expected-skills-sh-catalog-entries 89465 \
+            --expected-skills-sh-converted 89465 \
+            --expected-skill-pages 91464 \
+            --expected-agent-pages 467 \
+            --expected-mcp-pages 10790 \
+            --expected-harness-pages 207 \
+            --line-threshold 180 \
+            --max-stage-lines 40
+
+      - name: Validate README and docs stats
+        if: steps.resolve_graph_assets.outcome == 'success'
+        run: python src/update_repo_stats.py --check
+
+      - name: Static gates
+        run: |
+          python -m ruff check src hooks scripts
+          python -m mypy src
+
+      - name: Clean-host contract
+        run: python scripts/clean_host_contract.py --fast
+
+      - name: Release canary tests
+        run: |
+          python -m pytest -q --no-cov \
+            src/tests/test_ci_classifier.py \
+            src/tests/test_package_scaffold.py \
+            src/tests/test_clean_host_contract.py \
+            src/tests/test_alive_loop_e2e.py
+
+      - name: Build distributions
+        run: python -m build
+
+      - name: Check distributions
+        run: python -m twine check dist/*
+
+      - name: Check distribution contents
+        run: |
+          python - <<'PY'
+          import tarfile
+          import zipfile
+          from pathlib import Path
+
+          wheel = next(Path("dist").glob("*.whl"))
+          sdist = next(Path("dist").glob("*.tar.gz"))
+
+          with zipfile.ZipFile(wheel) as zf:
+              wheel_names = set(zf.namelist())
+          required = {"ctx/config.json", "ctx/skill-registry.json"}
+          missing = sorted(required - wheel_names)
+          if missing:
+              raise SystemExit(f"wheel missing packaged defaults: {missing}")
+
+          with tarfile.open(sdist, "r:gz") as tf:
+              names = set(tf.getnames())
+          forbidden = ("/src/tests/", "/.claude/", "/.a5c/")
+          leaked = sorted(name for name in names if any(part in name for part in forbidden))
+          if leaked:
+              raise SystemExit("sdist contains local/test artifacts:\n" + "\n".join(leaked[:20]))
+          print(f"checked distribution contents: {wheel.name}, {sdist.name}")
+          PY
+
+      - name: Smoke install wheel
+        run: |
+          python -m venv .venv-smoke
+          . .venv-smoke/bin/activate
+          python -m pip install --upgrade pip
+          wheel="$(python - <<'PY'
+          from pathlib import Path
+          print(next(Path("dist").glob("*.whl")))
+          PY
+          )"
+          python -m pip install "$wheel"
+          python -m pip check
+          python - <<'PY'
+          import tomllib
+          from importlib.metadata import entry_points, version
+
+          import ctx
+          import ctx_config
+
+          with open("pyproject.toml", "rb") as fh:
+              expected_scripts = set(tomllib.load(fh)["project"]["scripts"])
+
+          dist_version = version("claude-ctx")
+          if ctx.__version__ != dist_version:
+              raise SystemExit(
+                  f"ctx.__version__={ctx.__version__!r} != metadata {dist_version!r}"
+              )
+          if ctx_config.cfg.recommendation_top_k != 5:
+              raise SystemExit("packaged default config was not loaded")
+          scripts = {
+              ep.name: ep for ep in entry_points(group="console_scripts")
+              if ep.name == "ctx" or ep.name.startswith("ctx-")
+          }
+          missing = sorted(expected_scripts - set(scripts))
+          extra = sorted(set(scripts) - expected_scripts)
+          if missing or extra:
+              raise SystemExit(
+                  "wheel console-script surface mismatch\n"
+                  f"missing: {missing}\n"
+                  f"extra: {extra}"
+              )
+          failures = []
+          for ep in scripts.values():
+              try:
+                  ep.load()
+              except Exception as exc:
+                  failures.append(f"{ep.name}: {exc!r}")
+          if failures:
+              raise SystemExit("console script load failures:\n" + "\n".join(failures))
+          unsafe_help = {"ctx-mcp-server"}
+          safe_help = sorted(expected_scripts - unsafe_help)
+          with open("ctx-console-help.txt", "w", encoding="utf-8", newline="\n") as fh:
+              fh.write("\n".join(safe_help) + "\n")
+          print(
+              f"loaded {len(scripts)} ctx console scripts from wheel "
+              f"{dist_version}; help-smoke={len(safe_help)}"
+          )
+          PY
+          while IFS= read -r cmd; do
+            cmd="${cmd%$'\r'}"
+            [[ -z "$cmd" ]] && continue
+            echo "help smoke: $cmd"
+            "$cmd" --help >/dev/null
+          done < ctx-console-help.txt
+          python -m pip install "${wheel}[harness]"
+          python - <<'PY'
+          import litellm
+
+          print(f"harness extra import ok: litellm {getattr(litellm, '__version__', 'unknown')}")
+          PY
+
+      - name: Upload dist artifact
+        uses: actions/upload-artifact@v7
+        with:
+          name: dist
+          path: dist/
+
+      - name: Upload graph artifact bundle
+        if: steps.resolve_graph_assets.outcome == 'success'
+        uses: actions/upload-artifact@v7
+        with:
+          name: graph-release-assets
+          path: |
+            graph/wiki-graph.tar.gz
+            graph/wiki-graph-runtime.tar.gz
+            graph/skills-sh-catalog.json.gz
+            graph/communities.json
+            graph/entity-overlays.jsonl
+
+  release-assets:
+    name: Upload graph release assets
+    needs: build
+    if: github.event_name == 'push' && needs.build.outputs.graph_assets_available == 'true'
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write
+    steps:
+      - name: Download graph artifact bundle
+        uses: actions/download-artifact@v7
+        with:
+          name: graph-release-assets
+          path: graph-release-assets
+
+      - name: Upload graph assets to GitHub release
+        env:
+          GH_TOKEN: ${{ github.token }}
+          TAG_NAME: ${{ github.ref_name }}
+        run: |
+          gh release view "$TAG_NAME" --repo "$GITHUB_REPOSITORY" >/dev/null 2>&1 \
+            || gh release create "$TAG_NAME" \
+              --repo "$GITHUB_REPOSITORY" \
+              --title "$TAG_NAME" \
+              --notes "ctx $TAG_NAME"
+          gh release upload "$TAG_NAME" \
+            --repo "$GITHUB_REPOSITORY" \
+            graph-release-assets/wiki-graph.tar.gz \
+            graph-release-assets/wiki-graph-runtime.tar.gz \
+            graph-release-assets/skills-sh-catalog.json.gz \
+            graph-release-assets/communities.json \
+            graph-release-assets/entity-overlays.jsonl \
+            --clobber
+
+  publish:
+    name: Publish to PyPI
+    needs:
+      - build
+      - release-assets
+    if: ${{ always() && needs.build.result == 'success' && (needs.release-assets.result == 'success' || (github.event_name == 'workflow_dispatch' && github.event.inputs.repository == 'testpypi')) }}
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      id-token: write   # required for Trusted Publishing
+    environment:
+      name: pypi
+      url: https://pypi.org/project/claude-ctx/
+    steps:
+      - name: Download dist artifact
+        uses: actions/download-artifact@v7
+        with:
+          name: dist
+          path: dist/
+
+      - name: Publish to PyPI
+        if: github.event_name == 'push'
+        uses: pypa/gh-action-pypi-publish@release/v1
+
+      - name: Publish to TestPyPI
+        if: github.event_name == 'workflow_dispatch' && github.event.inputs.repository == 'testpypi'
+        uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          repository-url: https://test.pypi.org/legacy/

.github/workflows/test.yml

@@ -1,704 +1,710 @@
-name: Tests
-
-on:
-  push:
-    branches: [main]
-  pull_request:
-    branches: [main]
-
-env:
-  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
-
-permissions:
-  contents: read
-
-concurrency:
-  group: tests-${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
-
-jobs:
-  classify:
-    name: "Change classifier"
-    runs-on: ubuntu-latest
-    outputs:
-      browser_changed: ${{ steps.classify.outputs.browser_changed }}
-      ci_changed: ${{ steps.classify.outputs.ci_changed }}
-      docs_changed: ${{ steps.classify.outputs.docs_changed }}
-      docs_only: ${{ steps.classify.outputs.docs_only }}
-      graph_artifact_changed: ${{ steps.classify.outputs.graph_artifact_changed }}
-      graph_changed: ${{ steps.classify.outputs.graph_changed }}
-      graph_only: ${{ steps.classify.outputs.graph_only }}
-      package_changed: ${{ steps.classify.outputs.package_changed }}
-      similarity_changed: ${{ steps.classify.outputs.similarity_changed }}
-      source_changed: ${{ steps.classify.outputs.source_changed }}
-    steps:
-      - name: Checkout with full history
-        uses: actions/checkout@v5
-        with:
-          fetch-depth: 0
-          lfs: false
-
-      - name: Classify changed paths
-        id: classify
-        shell: bash
-        env:
-          BEFORE_SHA: ${{ github.event.before }}
-          EVENT_NAME: ${{ github.event_name }}
-          HEAD_SHA: ${{ github.sha }}
-          PR_BASE_SHA: ${{ github.event.pull_request.base.sha }}
-        run: |
-          set -euo pipefail
-          BASE="$BEFORE_SHA"
-          if [[ "$EVENT_NAME" == "pull_request" && -n "$PR_BASE_SHA" ]]; then
-            BASE="$PR_BASE_SHA"
-          fi
-          if [[ -z "$BASE" || "$BASE" =~ ^0+$ ]]; then
-            BASE="$(git rev-list --max-parents=0 "$HEAD_SHA")"
-          fi
-          if git cat-file -e "$BASE^{commit}" 2>/dev/null; then
-            git diff --name-only "$BASE" "$HEAD_SHA" > changed-files.txt
-          elif git cat-file -e "$HEAD_SHA^" 2>/dev/null; then
-            git diff --name-only "$HEAD_SHA^" "$HEAD_SHA" > changed-files.txt
-          else
-            git diff-tree --no-commit-id --name-only -r "$HEAD_SHA" > changed-files.txt
-          fi
-
-          python scripts/ci_classifier.py changed-files.txt
-
-  static:
-    name: "Static quality gates"
-    needs: classify
-    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python 3.11
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.11"
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Install dependencies
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install ".[dev]"
-
-      - name: Run static quality gates
-        run: python -m ruff check src hooks scripts && python -m mypy src && python -m pip check
-
-  unit-linux:
-    name: "unit-linux"
-    needs: classify
-    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python 3.11
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.11"
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Install dependencies
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install ".[dev]"
-
-      - name: Run tests with coverage gate
-        # Bump this number as the coverage sprint ships - never lower it.
-        run: pytest -q -m "not browser and not integration" --cov=src --cov-report=term-missing --cov-fail-under=40
-
-      - name: Upload coverage artifact
-        uses: actions/upload-artifact@v7
-        with:
-          name: coverage
-          path: .coverage
-          include-hidden-files: true
-          retention-days: 7
-
-  similarity-integration:
-    name: "Similarity precision/recall"
-    needs: classify
-    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.similarity_changed == 'true') }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python 3.11
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.11"
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Install embedding dependencies
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install ".[dev,embeddings]"
-
-      - name: Run similarity precision/recall gate
-        env:
-          CTX_REQUIRE_SIMILARITY_EVAL: "1"
-        run: pytest -q --no-cov -m integration src/tests/test_similarity_precision_recall.py
-
-  test:
-    name: pytest (${{ matrix.os }} / py${{ matrix.python-version }})
-    if: ${{ github.event_name != 'pull_request' }}
-    runs-on: ${{ matrix.os }}
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - os: ubuntu-latest
-            python-version: "3.12"
-          - os: windows-latest
-            python-version: "3.11"
-          - os: windows-latest
-            python-version: "3.12"
-          - os: macos-latest
-            python-version: "3.11"
-          - os: macos-latest
-            python-version: "3.12"
-
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v6
-        with:
-          python-version: ${{ matrix.python-version }}
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Install dependencies
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install ".[dev]"
-
-      - name: Run tests without coverage
-        run: pytest -q -m "not browser and not integration" --no-cov
-
-  contract-compat:
-    name: "Contract compatibility (${{ matrix.os }})"
-    needs: classify
-    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
-    runs-on: ${{ matrix.os }}
-    strategy:
-      fail-fast: false
-      matrix:
-        os: [windows-latest, macos-latest]
-
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python 3.12
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.12"
-
-      - name: Install dependencies
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install ".[dev]"
-
-      - name: Run cross-OS contract tests
-        run: python -m pytest -q --no-cov src/tests/test_clean_host_contract.py src/tests/test_package_scaffold.py
-
-  e2e-canary:
-    name: "A-Z alive-loop E2E canary"
-    needs: classify
-    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python 3.11
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.11"
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Install dependencies
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install ".[dev]"
-
-      - name: Run E2E + fuzz canary suite
-        # These two files pin the critical-path invariants:
-        #   test_alive_loop_e2e.py: A-Z user journey (signals → bundle →
-        #     install → unload → purge) — any handoff regression trips it.
-        #   test_fuzz_yaml_rendering.py: property-based YAML injection
-        #     coverage on install/enrich render_scalar.
-        # They live in the main matrix job already, but pulling them
-        # forward as a separate fail-fast canary surfaces the regression
-        # in seconds instead of waiting for the full test run.
-        run: |
-          pytest -q --no-cov \
-            src/tests/test_alive_loop_e2e.py \
-            src/tests/test_fuzz_yaml_rendering.py
-
-  docs-check:
-    name: "Docs strict build"
-    needs: classify
-    if: ${{ github.event_name == 'pull_request' && needs.classify.outputs.docs_changed == 'true' }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python 3.11
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.11"
-          cache: pip
-          cache-dependency-path: requirements-docs.txt
-
-      - name: Build docs strictly
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install -r requirements-docs.txt
-          python -m mkdocs build --strict
-
-  graph-check:
-    name: "Graph artifact check"
-    needs: classify
-    if: ${{ github.event_name == 'pull_request' && needs.classify.outputs.graph_artifact_changed == 'true' }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout graph artifacts
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Resolve graph LFS artifacts
-        shell: bash
-        env:
-          GH_TOKEN: ${{ github.token }}
-        run: |
-          set -euo pipefail
-          git lfs install --local
-          if git lfs pull --include="graph/wiki-graph.tar.gz,graph/wiki-graph-runtime.tar.gz" --exclude=""; then
-            exit 0
-          fi
-          echo "Git LFS download failed; trying matching prior release asset."
-          python - <<'PY'
-          import hashlib
-          import json
-          import os
-          from pathlib import Path
-          import subprocess
-          import time
-          import urllib.request
-
-          repo = os.environ["GITHUB_REPOSITORY"]
-          release_asset_wait_seconds = 300
-          release_asset_poll_seconds = 10
-
-          def load_releases() -> list[dict]:
-              return json.loads(subprocess.check_output(
-                  ["gh", "api", f"repos/{repo}/releases?per_page=50"],
-                  text=True,
-              ))
-
-          def hydrate_from_release(path_name: str, hydrated_min_size: int) -> None:
-              graph_tar = Path(path_name)
-              pointer = graph_tar.read_text(encoding="utf-8", errors="replace")
-              expected_oid = ""
-              expected_size = 0
-              for line in pointer.splitlines():
-                  if line.startswith("oid sha256:"):
-                      expected_oid = line.split(":", 1)[1].strip()
-                  elif line.startswith("size "):
-                      expected_size = int(line.split(" ", 1)[1].strip())
-              if not expected_oid:
-                  if graph_tar.stat().st_size > hydrated_min_size:
-                      print(f"{graph_tar} is already hydrated")
-                      return
-                  raise SystemExit(f"{path_name} is neither hydrated nor an LFS pointer")
-
-              deadline = time.monotonic() + release_asset_wait_seconds
-              while True:
-                  candidates = []
-                  for release in load_releases():
-                      tag_name = str(release.get("tag_name") or "")
-                      is_graph_cache = tag_name.startswith("graph-artifacts-")
-                      if release.get("draft") or (
-                          release.get("prerelease") and not is_graph_cache
-                      ):
-                          continue
-                      for asset in release.get("assets", []):
-                          if asset.get("name") != graph_tar.name:
-                              continue
-                          digest = str(asset.get("digest") or "")
-                          size = int(asset.get("size") or 0)
-                          if size != expected_size:
-                              continue
-                          if digest and digest != f"sha256:{expected_oid}":
-                              continue
-                          candidates.append((tag_name, asset))
-
-                  if candidates:
-                      break
-                  if time.monotonic() >= deadline:
-                      raise SystemExit(
-                          f"No previous release asset matches {path_name} "
-                          f"sha256:{expected_oid} size:{expected_size}"
-                      )
-                  print(
-                      f"Waiting for matching release asset {graph_tar.name} "
-                      f"sha256:{expected_oid} size:{expected_size}"
-                  )
-                  time.sleep(release_asset_poll_seconds)
-
-              source_tag, asset = candidates[0]
-              tmp = graph_tar.with_name(f"{graph_tar.name}.download")
-              sha = hashlib.sha256()
-              total = 0
-              with urllib.request.urlopen(asset["browser_download_url"], timeout=300) as resp:  # noqa: S310
-                  with tmp.open("wb") as fh:
-                      while True:
-                          chunk = resp.read(1024 * 1024)
-                          if not chunk:
-                              break
-                          sha.update(chunk)
-                          total += len(chunk)
-                          fh.write(chunk)
-              actual_oid = sha.hexdigest()
-              if actual_oid != expected_oid or total != expected_size:
-                  tmp.unlink(missing_ok=True)
-                  raise SystemExit(
-                      f"Downloaded {path_name} does not match LFS pointer: "
-                      f"sha256:{actual_oid} size:{total}"
-                  )
-              tmp.replace(graph_tar)
-              print(
-                  f"Hydrated {path_name} from {source_tag} release asset "
-                  f"sha256:{actual_oid} size:{total}"
-              )
-
-          hydrate_from_release("graph/wiki-graph.tar.gz", 100_000_000)
-          hydrate_from_release("graph/wiki-graph-runtime.tar.gz", 10_000_000)
-          PY
-
-      - name: Validate shipped graph artifacts
-        run: |
-          python src/validate_graph_artifacts.py \
-            --graph-dir graph \
-            --deep \
-            --min-nodes 100000 \
-            --min-edges 2000000 \
-            --min-skills-sh-nodes 89000 \
-            --min-semantic-edges 1000000 \
-            --expected-nodes 102928 \
-            --expected-edges 2913960 \
-            --expected-semantic-edges 1683193 \
-            --expected-harness-nodes 207 \
-            --expected-skills-sh-nodes 89471 \
-            --expected-skills-sh-catalog-entries 89465 \
-            --expected-skills-sh-converted 89465 \
-            --expected-skill-pages 91464 \
-            --expected-agent-pages 467 \
-            --expected-mcp-pages 10790 \
-            --expected-harness-pages 207 \
-            --line-threshold 180 \
-            --max-stage-lines 40
-
-  browser-security:
-    name: "Browser monitor security"
-    needs: classify
-    if: ${{ github.event_name == 'push' || needs.classify.outputs.browser_changed == 'true' }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python 3.12
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.12"
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Install browser Python dependencies
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install ".[dev,browser]"
-
-      - name: Detect Playwright version
-        id: playwright
-        run: |
-          python - <<'PY' >> "$GITHUB_OUTPUT"
-          from importlib.metadata import version
-
-          print(f"version={version('playwright')}")
-          PY
-
-      - name: Cache Playwright browsers
-        uses: actions/cache@v4
-        with:
-          path: ~/.cache/ms-playwright
-          key: playwright-${{ runner.os }}-${{ steps.playwright.outputs.version }}
-
-      - name: Install browser runtime
-        run: |
-          python -m playwright install-deps chromium
-          python -m playwright install chromium
-
-      - name: Run browser security tests
-        run: pytest -q --no-cov -m browser src/tests/test_ctx_monitor_browser.py
-
-  package-build:
-    name: "Build wheel package"
-    needs: classify
-    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python 3.12
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.12"
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Build and inspect wheel
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install build twine
-          python -m build
-          python -m twine check dist/*
-          python - <<'PY'
-          from pathlib import Path
-
-          wheels = sorted(Path("dist").glob("*.whl"))
-          if len(wheels) != 1:
-              raise SystemExit(f"expected exactly one wheel, found {len(wheels)}")
-          wheel = wheels[0]
-          if not wheel.name.endswith("-py3-none-any.whl"):
-              message = f"expected a universal pure-Python wheel, got {wheel.name}"
-              raise SystemExit(message)
-          print(f"built package-smoke wheel: {wheel.name}")
-          PY
-
-      - name: Upload wheel artifact
-        uses: actions/upload-artifact@v7
-        with:
-          name: package-smoke-wheel
-          path: dist/*.whl
-          retention-days: 7
-
-  package-smoke:
-    name: "Wheel package smoke (${{ matrix.os }})"
-    needs:
-      - classify
-      - package-build
-    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
-    runs-on: ${{ matrix.os }}
-    strategy:
-      fail-fast: false
-      matrix:
-        os: [ubuntu-latest, windows-latest, macos-latest]
-    defaults:
-      run:
-        shell: bash
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python 3.12
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.12"
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Download wheel artifact
-        uses: actions/download-artifact@v7
-        with:
-          name: package-smoke-wheel
-          path: dist
-
-      - name: Install wheel in clean venv
-        run: |
-          python -m venv .venv-smoke
-          if [[ "$RUNNER_OS" == "Windows" ]]; then
-            . .venv-smoke/Scripts/activate
-          else
-            . .venv-smoke/bin/activate
-          fi
-          python -m pip install --upgrade pip
-          python -m pip install dist/*.whl
-          python -m pip check
-          python - <<'PY'
-          import tomllib
-          from importlib.metadata import entry_points, version
-
-          import ctx
-
-          with open("pyproject.toml", "rb") as fh:
-              expected_scripts = set(tomllib.load(fh)["project"]["scripts"])
-
-          dist_version = version("claude-ctx")
-          if ctx.__version__ != dist_version:
-              raise SystemExit(
-                  f"ctx.__version__={ctx.__version__!r} != metadata {dist_version!r}"
-              )
-
-          script_eps = {
-              ep.name: ep
-              for ep in entry_points(group="console_scripts")
-              if ep.name == "ctx" or ep.name.startswith("ctx-")
-          }
-          missing = sorted(expected_scripts - set(script_eps))
-          extra = sorted(set(script_eps) - expected_scripts)
-          if missing or extra:
-              raise SystemExit(
-                  "wheel console-script surface mismatch\n"
-                  f"missing: {missing}\n"
-                  f"extra: {extra}"
-              )
-          failures = []
-          for ep in script_eps.values():
-              try:
-                  ep.load()
-              except Exception as exc:
-                  failures.append(f"{ep.name}: {exc!r}")
-          if failures:
-              raise SystemExit("console script load failures:\n" + "\n".join(failures))
-
-          unsafe_help = {"ctx-mcp-server"}
-          safe_help = sorted(expected_scripts - unsafe_help)
-          with open(
-              "ctx-console-help.txt", "w", encoding="utf-8", newline="\n"
-          ) as fh:
-              fh.write("\n".join(safe_help) + "\n")
-          print(
-              f"loaded {len(script_eps)} ctx console scripts from wheel "
-              f"{dist_version}; help-smoke={len(safe_help)}"
-          )
-          PY
-          while IFS= read -r cmd; do
-            cmd="${cmd%$'\r'}"
-            [[ -z "$cmd" ]] && continue
-            echo "help smoke: $cmd"
-            "$cmd" --help >/dev/null
-          done < ctx-console-help.txt
-
-  clean-host-contract:
-    name: "Clean host contract"
-    needs: classify
-    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
-    runs-on: ubuntu-latest
-    timeout-minutes: 25
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python 3.12
-        uses: actions/setup-python@v6
-        with:
-          python-version: "3.12"
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Upgrade pip
-        run: python -m pip install --upgrade pip
-
-      - name: Run clean-host contract
-        run: python scripts/clean_host_contract.py --fast
-
-  no-test-no-merge:
-    name: "Contract changes must touch tests"
-    needs: classify
-    runs-on: ubuntu-latest
-    if: ${{ github.event_name == 'pull_request' && needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true' }}
-    steps:
-      - name: Checkout with full history
-        uses: actions/checkout@v5
-        with:
-          fetch-depth: 0
-          lfs: false
-
-      - name: Enforce test-coverage-per-PR policy
-        # Policy: any product or CI/package contract change
-        # must be accompanied by at least one test file change in the
-        # same PR. Exemptions:
-        #   - Pure docs / comment changes (matched by a trivial heuristic)
-        #   - Config JSON (src/config.json) — schema changes get tests
-        #     for the accessor, not the JSON itself.
-        #   - Changes ONLY inside __pycache__, .pyc files, etc.
-        run: |
-          set -euo pipefail
-          BASE="${{ github.event.pull_request.base.sha }}"
-          HEAD="${{ github.event.pull_request.head.sha }}"
-          LABELS='${{ toJson(github.event.pull_request.labels.*.name) }}'
-          python scripts/ci_no_test_policy.py \
-            --base "$BASE" \
-            --head "$HEAD" \
-            --labels-json "$LABELS"
-
-  ci-required:
-    name: "CI required"
-    runs-on: ubuntu-latest
-    needs:
-      - classify
-      - static
-      - unit-linux
-      - similarity-integration
-      - test
-      - contract-compat
-      - e2e-canary
-      - docs-check
-      - graph-check
-      - browser-security
-      - package-smoke
-      - package-build
-      - clean-host-contract
-      - no-test-no-merge
-    if: ${{ always() }}
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Check required job results
-        env:
-          EVENT_NAME: ${{ github.event_name }}
-          NEEDS_JSON: ${{ toJson(needs) }}
-        run: python scripts/ci_required.py
+name: Tests
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+env:
+  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
+
+permissions:
+  contents: read
+
+concurrency:
+  group: tests-${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  classify:
+    name: "Change classifier"
+    runs-on: ubuntu-latest
+    outputs:
+      browser_changed: ${{ steps.classify.outputs.browser_changed }}
+      ci_changed: ${{ steps.classify.outputs.ci_changed }}
+      docs_changed: ${{ steps.classify.outputs.docs_changed }}
+      docs_only: ${{ steps.classify.outputs.docs_only }}
+      graph_artifact_changed: ${{ steps.classify.outputs.graph_artifact_changed }}
+      graph_changed: ${{ steps.classify.outputs.graph_changed }}
+      graph_only: ${{ steps.classify.outputs.graph_only }}
+      package_changed: ${{ steps.classify.outputs.package_changed }}
+      similarity_changed: ${{ steps.classify.outputs.similarity_changed }}
+      source_changed: ${{ steps.classify.outputs.source_changed }}
+    steps:
+      - name: Checkout with full history
+        uses: actions/checkout@v5
+        with:
+          fetch-depth: 0
+          lfs: false
+
+      - name: Classify changed paths
+        id: classify
+        shell: bash
+        env:
+          BEFORE_SHA: ${{ github.event.before }}
+          EVENT_NAME: ${{ github.event_name }}
+          HEAD_SHA: ${{ github.sha }}
+          PR_BASE_SHA: ${{ github.event.pull_request.base.sha }}
+        run: |
+          set -euo pipefail
+          BASE="$BEFORE_SHA"
+          if [[ "$EVENT_NAME" == "pull_request" && -n "$PR_BASE_SHA" ]]; then
+            BASE="$PR_BASE_SHA"
+          fi
+          if [[ -z "$BASE" || "$BASE" =~ ^0+$ ]]; then
+            BASE="$(git rev-list --max-parents=0 "$HEAD_SHA")"
+          fi
+          if git cat-file -e "$BASE^{commit}" 2>/dev/null; then
+            git diff --name-only "$BASE" "$HEAD_SHA" > changed-files.txt
+          elif git cat-file -e "$HEAD_SHA^" 2>/dev/null; then
+            git diff --name-only "$HEAD_SHA^" "$HEAD_SHA" > changed-files.txt
+          else
+            git diff-tree --no-commit-id --name-only -r "$HEAD_SHA" > changed-files.txt
+          fi
+
+          python scripts/ci_classifier.py changed-files.txt
+
+  static:
+    name: "Static quality gates"
+    needs: classify
+    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python 3.11
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install ".[dev]"
+
+      - name: Run static quality gates
+        run: python -m ruff check src hooks scripts && python -m mypy src && python -m pip check
+
+  unit-linux:
+    name: "unit-linux"
+    needs: classify
+    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python 3.11
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install ".[dev]"
+
+      - name: Run tests with coverage gate
+        # Bump this number as the coverage sprint ships - never lower it.
+        run: pytest -q -m "not browser and not integration" --cov=src --cov-report=term-missing --cov-fail-under=40
+
+      - name: Upload coverage artifact
+        uses: actions/upload-artifact@v7
+        with:
+          name: coverage
+          path: .coverage
+          include-hidden-files: true
+          retention-days: 7
+
+  similarity-integration:
+    name: "Similarity precision/recall"
+    needs: classify
+    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.similarity_changed == 'true') }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python 3.11
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Install embedding dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install ".[dev,embeddings]"
+
+      - name: Run similarity precision/recall gate
+        env:
+          CTX_REQUIRE_SIMILARITY_EVAL: "1"
+        run: pytest -q --no-cov -m integration src/tests/test_similarity_precision_recall.py
+
+  test:
+    name: pytest (${{ matrix.os }} / py${{ matrix.python-version }})
+    if: ${{ github.event_name != 'pull_request' }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - os: ubuntu-latest
+            python-version: "3.12"
+          - os: windows-latest
+            python-version: "3.11"
+          - os: windows-latest
+            python-version: "3.12"
+          - os: macos-latest
+            python-version: "3.11"
+          - os: macos-latest
+            python-version: "3.12"
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v6
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install ".[dev]"
+
+      - name: Run tests without coverage
+        run: pytest -q -m "not browser and not integration" --no-cov
+
+  contract-compat:
+    name: "Contract compatibility (${{ matrix.os }})"
+    needs: classify
+    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [windows-latest, macos-latest]
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python 3.12
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.12"
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install ".[dev]"
+
+      - name: Run cross-OS contract tests
+        run: python -m pytest -q --no-cov src/tests/test_clean_host_contract.py src/tests/test_package_scaffold.py
+
+  e2e-canary:
+    name: "A-Z alive-loop E2E canary"
+    needs: classify
+    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python 3.11
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install ".[dev]"
+
+      - name: Run E2E + fuzz canary suite
+        # These two files pin the critical-path invariants:
+        #   test_alive_loop_e2e.py: A-Z user journey (signals → bundle →
+        #     install → unload → purge) — any handoff regression trips it.
+        #   test_fuzz_yaml_rendering.py: property-based YAML injection
+        #     coverage on install/enrich render_scalar.
+        # They live in the main matrix job already, but pulling them
+        # forward as a separate fail-fast canary surfaces the regression
+        # in seconds instead of waiting for the full test run.
+        run: |
+          pytest -q --no-cov \
+            src/tests/test_alive_loop_e2e.py \
+            src/tests/test_fuzz_yaml_rendering.py
+
+  docs-check:
+    name: "Docs strict build"
+    needs: classify
+    if: ${{ github.event_name == 'pull_request' && needs.classify.outputs.docs_changed == 'true' }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python 3.11
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.11"
+          cache: pip
+          cache-dependency-path: requirements-docs.txt
+
+      - name: Build docs strictly
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install -r requirements-docs.txt
+          python -m mkdocs build --strict
+
+  graph-check:
+    name: "Graph artifact check"
+    needs: classify
+    if: ${{ github.event_name == 'pull_request' && needs.classify.outputs.graph_artifact_changed == 'true' }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout graph artifacts
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Resolve graph artifacts from release assets
+        shell: bash
+        env:
+          GH_TOKEN: ${{ github.token }}
+        run: |
+          set -euo pipefail
+          echo "Resolving graph artifacts from matching release assets to avoid Git LFS bandwidth."
+          python - <<'PY'
+          import hashlib
+          import json
+          import os
+          from pathlib import Path
+          import subprocess
+          import time
+          import urllib.request
+
+          repo = os.environ["GITHUB_REPOSITORY"]
+          release_asset_wait_seconds = 300
+          release_asset_poll_seconds = 10
+          expected_graph_assets = {
+              "graph/wiki-graph.tar.gz": {
+                  "sha256": "91b30795e7d200cf31a62a8749969d12658f5f74636d2de06d6b2b24b393c12f",
+                  "size": 457545675,
+              },
+              "graph/wiki-graph-runtime.tar.gz": {
+                  "sha256": "334fb19bace3fd6e4b92087850f17297fb248032957d123f3f1432dfde2e36c0",
+                  "size": 175773376,
+              },
+          }
+
+          def load_releases() -> list[dict]:
+              return json.loads(subprocess.check_output(
+                  ["gh", "api", f"repos/{repo}/releases?per_page=50"],
+                  text=True,
+              ))
+
+          def hydrate_from_release(path_name: str, hydrated_min_size: int) -> None:
+              graph_tar = Path(path_name)
+              fallback = expected_graph_assets[path_name]
+              expected_oid = fallback["sha256"]
+              expected_size = int(fallback["size"])
+              if graph_tar.exists():
+                  pointer = graph_tar.read_text(encoding="utf-8", errors="replace")
+                  for line in pointer.splitlines():
+                      if line.startswith("oid sha256:"):
+                          expected_oid = line.split(":", 1)[1].strip()
+                      elif line.startswith("size "):
+                          expected_size = int(line.split(" ", 1)[1].strip())
+                  if not pointer.startswith("version https://git-lfs.github.com/spec/v1") and graph_tar.stat().st_size > hydrated_min_size:
+                      print(f"{graph_tar} is already hydrated")
+                      return
+
+              deadline = time.monotonic() + release_asset_wait_seconds
+              while True:
+                  candidates = []
+                  for release in load_releases():
+                      tag_name = str(release.get("tag_name") or "")
+                      is_graph_cache = tag_name.startswith("graph-artifacts-")
+                      if release.get("draft") or (
+                          release.get("prerelease") and not is_graph_cache
+                      ):
+                          continue
+                      for asset in release.get("assets", []):
+                          if asset.get("name") != graph_tar.name:
+                              continue
+                          digest = str(asset.get("digest") or "")
+                          size = int(asset.get("size") or 0)
+                          if size != expected_size:
+                              continue
+                          if digest and digest != f"sha256:{expected_oid}":
+                              continue
+                          candidates.append((tag_name, asset))
+
+                  if candidates:
+                      break
+                  if time.monotonic() >= deadline:
+                      raise SystemExit(
+                          f"No previous release asset matches {path_name} "
+                          f"sha256:{expected_oid} size:{expected_size}"
+                      )
+                  print(
+                      f"Waiting for matching release asset {graph_tar.name} "
+                      f"sha256:{expected_oid} size:{expected_size}"
+                  )
+                  time.sleep(release_asset_poll_seconds)
+
+              source_tag, asset = candidates[0]
+              tmp = graph_tar.with_name(f"{graph_tar.name}.download")
+              sha = hashlib.sha256()
+              total = 0
+              with urllib.request.urlopen(asset["browser_download_url"], timeout=300) as resp:  # noqa: S310
+                  with tmp.open("wb") as fh:
+                      while True:
+                          chunk = resp.read(1024 * 1024)
+                          if not chunk:
+                              break
+                          sha.update(chunk)
+                          total += len(chunk)
+                          fh.write(chunk)
+              actual_oid = sha.hexdigest()
+              if actual_oid != expected_oid or total != expected_size:
+                  tmp.unlink(missing_ok=True)
+                  raise SystemExit(
+                      f"Downloaded {path_name} does not match LFS pointer: "
+                      f"sha256:{actual_oid} size:{total}"
+                  )
+              tmp.replace(graph_tar)
+              print(
+                  f"Hydrated {path_name} from {source_tag} release asset "
+                  f"sha256:{actual_oid} size:{total}"
+              )
+
+          hydrate_from_release("graph/wiki-graph.tar.gz", 100_000_000)
+          hydrate_from_release("graph/wiki-graph-runtime.tar.gz", 10_000_000)
+          PY
+
+      - name: Validate shipped graph artifacts
+        run: |
+          python src/validate_graph_artifacts.py \
+            --graph-dir graph \
+            --deep \
+            --min-nodes 100000 \
+            --min-edges 2000000 \
+            --min-skills-sh-nodes 89000 \
+            --min-semantic-edges 1000000 \
+            --expected-nodes 102928 \
+            --expected-edges 2913960 \
+            --expected-semantic-edges 1683193 \
+            --expected-harness-nodes 207 \
+            --expected-skills-sh-nodes 89471 \
+            --expected-skills-sh-catalog-entries 89465 \
+            --expected-skills-sh-converted 89465 \
+            --expected-skill-pages 91464 \
+            --expected-agent-pages 467 \
+            --expected-mcp-pages 10790 \
+            --expected-harness-pages 207 \
+            --line-threshold 180 \
+            --max-stage-lines 40
+
+  browser-security:
+    name: "Browser monitor security"
+    needs: classify
+    if: ${{ github.event_name == 'push' || needs.classify.outputs.browser_changed == 'true' }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python 3.12
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.12"
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Install browser Python dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install ".[dev,browser]"
+
+      - name: Detect Playwright version
+        id: playwright
+        run: |
+          python - <<'PY' >> "$GITHUB_OUTPUT"
+          from importlib.metadata import version
+
+          print(f"version={version('playwright')}")
+          PY
+
+      - name: Cache Playwright browsers
+        uses: actions/cache@v4
+        with:
+          path: ~/.cache/ms-playwright
+          key: playwright-${{ runner.os }}-${{ steps.playwright.outputs.version }}
+
+      - name: Install browser runtime
+        run: |
+          python -m playwright install-deps chromium
+          python -m playwright install chromium
+
+      - name: Run browser security tests
+        run: pytest -q --no-cov -m browser src/tests/test_ctx_monitor_browser.py
+
+  package-build:
+    name: "Build wheel package"
+    needs: classify
+    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python 3.12
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.12"
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Build and inspect wheel
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install build twine
+          python -m build
+          python -m twine check dist/*
+          python - <<'PY'
+          from pathlib import Path
+
+          wheels = sorted(Path("dist").glob("*.whl"))
+          if len(wheels) != 1:
+              raise SystemExit(f"expected exactly one wheel, found {len(wheels)}")
+          wheel = wheels[0]
+          if not wheel.name.endswith("-py3-none-any.whl"):
+              message = f"expected a universal pure-Python wheel, got {wheel.name}"
+              raise SystemExit(message)
+          print(f"built package-smoke wheel: {wheel.name}")
+          PY
+
+      - name: Upload wheel artifact
+        uses: actions/upload-artifact@v7
+        with:
+          name: package-smoke-wheel
+          path: dist/*.whl
+          retention-days: 7
+
+  package-smoke:
+    name: "Wheel package smoke (${{ matrix.os }})"
+    needs:
+      - classify
+      - package-build
+    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest, windows-latest, macos-latest]
+    defaults:
+      run:
+        shell: bash
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python 3.12
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.12"
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Download wheel artifact
+        uses: actions/download-artifact@v7
+        with:
+          name: package-smoke-wheel
+          path: dist
+
+      - name: Install wheel in clean venv
+        run: |
+          python -m venv .venv-smoke
+          if [[ "$RUNNER_OS" == "Windows" ]]; then
+            . .venv-smoke/Scripts/activate
+          else
+            . .venv-smoke/bin/activate
+          fi
+          python -m pip install --upgrade pip
+          python -m pip install dist/*.whl
+          python -m pip check
+          python - <<'PY'
+          import tomllib
+          from importlib.metadata import entry_points, version
+
+          import ctx
+
+          with open("pyproject.toml", "rb") as fh:
+              expected_scripts = set(tomllib.load(fh)["project"]["scripts"])
+
+          dist_version = version("claude-ctx")
+          if ctx.__version__ != dist_version:
+              raise SystemExit(
+                  f"ctx.__version__={ctx.__version__!r} != metadata {dist_version!r}"
+              )
+
+          script_eps = {
+              ep.name: ep
+              for ep in entry_points(group="console_scripts")
+              if ep.name == "ctx" or ep.name.startswith("ctx-")
+          }
+          missing = sorted(expected_scripts - set(script_eps))
+          extra = sorted(set(script_eps) - expected_scripts)
+          if missing or extra:
+              raise SystemExit(
+                  "wheel console-script surface mismatch\n"
+                  f"missing: {missing}\n"
+                  f"extra: {extra}"
+              )
+          failures = []
+          for ep in script_eps.values():
+              try:
+                  ep.load()
+              except Exception as exc:
+                  failures.append(f"{ep.name}: {exc!r}")
+          if failures:
+              raise SystemExit("console script load failures:\n" + "\n".join(failures))
+
+          unsafe_help = {"ctx-mcp-server"}
+          safe_help = sorted(expected_scripts - unsafe_help)
+          with open(
+              "ctx-console-help.txt", "w", encoding="utf-8", newline="\n"
+          ) as fh:
+              fh.write("\n".join(safe_help) + "\n")
+          print(
+              f"loaded {len(script_eps)} ctx console scripts from wheel "
+              f"{dist_version}; help-smoke={len(safe_help)}"
+          )
+          PY
+          while IFS= read -r cmd; do
+            cmd="${cmd%$'\r'}"
+            [[ -z "$cmd" ]] && continue
+            echo "help smoke: $cmd"
+            "$cmd" --help >/dev/null
+          done < ctx-console-help.txt
+
+  clean-host-contract:
+    name: "Clean host contract"
+    needs: classify
+    if: ${{ github.event_name != 'pull_request' || (needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true') }}
+    runs-on: ubuntu-latest
+    timeout-minutes: 25
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python 3.12
+        uses: actions/setup-python@v6
+        with:
+          python-version: "3.12"
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Upgrade pip
+        run: python -m pip install --upgrade pip
+
+      - name: Run clean-host contract
+        run: python scripts/clean_host_contract.py --fast
+
+  no-test-no-merge:
+    name: "Contract changes must touch tests"
+    needs: classify
+    runs-on: ubuntu-latest
+    if: ${{ github.event_name == 'pull_request' && needs.classify.outputs.docs_only != 'true' && needs.classify.outputs.graph_only != 'true' }}
+    steps:
+      - name: Checkout with full history
+        uses: actions/checkout@v5
+        with:
+          fetch-depth: 0
+          lfs: false
+
+      - name: Enforce test-coverage-per-PR policy
+        # Policy: any product or CI/package contract change
+        # must be accompanied by at least one test file change in the
+        # same PR. Exemptions:
+        #   - Pure docs / comment changes (matched by a trivial heuristic)
+        #   - Config JSON (src/config.json) — schema changes get tests
+        #     for the accessor, not the JSON itself.
+        #   - Changes ONLY inside __pycache__, .pyc files, etc.
+        run: |
+          set -euo pipefail
+          BASE="${{ github.event.pull_request.base.sha }}"
+          HEAD="${{ github.event.pull_request.head.sha }}"
+          LABELS='${{ toJson(github.event.pull_request.labels.*.name) }}'
+          python scripts/ci_no_test_policy.py \
+            --base "$BASE" \
+            --head "$HEAD" \
+            --labels-json "$LABELS"
+
+  ci-required:
+    name: "CI required"
+    runs-on: ubuntu-latest
+    needs:
+      - classify
+      - static
+      - unit-linux
+      - similarity-integration
+      - test
+      - contract-compat
+      - e2e-canary
+      - docs-check
+      - graph-check
+      - browser-security
+      - package-smoke
+      - package-build
+      - clean-host-contract
+      - no-test-no-merge
+    if: ${{ always() }}
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Check required job results
+        env:
+          EVENT_NAME: ${{ github.event_name }}
+          NEEDS_JSON: ${{ toJson(needs) }}
+        run: python scripts/ci_required.py

.github/workflows/xdist-experiment.yml

@@ -1,90 +1,90 @@
-name: Xdist experiment
-
-on:
-  workflow_dispatch:
-  pull_request:
-    branches: [main]
-    paths:
-      - ".github/workflows/xdist-experiment.yml"
-
-env:
-  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
-
-concurrency:
-  group: xdist-${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
-  cancel-in-progress: true
-
-permissions:
-  contents: read
-
-jobs:
-  xdist:
-    name: xdist (${{ matrix.os }} / py${{ matrix.python-version }})
-    runs-on: ${{ matrix.os }}
-    strategy:
-      fail-fast: false
-      matrix:
-        os: [ubuntu-latest, windows-latest, macos-latest]
-        python-version: ["3.12"]
-    defaults:
-      run:
-        shell: bash
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v5
-        with:
-          lfs: false
-
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v6
-        with:
-          python-version: ${{ matrix.python-version }}
-          cache: pip
-          cache-dependency-path: pyproject.toml
-
-      - name: Install experiment dependencies
-        run: |
-          python -m pip install --upgrade pip
-          python -m pip install ".[dev]" pytest-xdist
-
-      - name: Run xdist experiment
-        run: |
-          set +e
-          start="$(python - <<'PY'
-          import time
-          print(time.monotonic())
-          PY
-          )"
-          python -m pytest -q -m "not browser" --no-cov \
-            -n 2 \
-            --dist=loadfile \
-            --max-worker-restart=0 \
-            --durations=25
-          status="$?"
-          end="$(python - <<'PY'
-          import time
-          print(time.monotonic())
-          PY
-          )"
-          elapsed="$(python - "$start" "$end" <<'PY'
-          import sys
-          start = float(sys.argv[1])
-          end = float(sys.argv[2])
-          print(f"{end - start:.1f}")
-          PY
-          )"
-
-          {
-            echo "## pytest-xdist experiment"
-            echo ""
-            echo "- OS: $RUNNER_OS"
-            echo "- Python: ${{ matrix.python-version }}"
-            echo "- Command: \`python -m pytest -q -m \"not browser\" --no-cov -n 2 --dist=loadfile --max-worker-restart=0 --durations=25\`"
-            echo "- Exit code: $status"
-            echo "- Elapsed seconds: $elapsed"
-          } >> "$GITHUB_STEP_SUMMARY"
-
-          if [ "$status" -ne 0 ]; then
-            echo "::warning::xdist experiment failed on $RUNNER_OS with exit code $status"
-          fi
-          exit "$status"
+name: Xdist experiment
+
+on:
+  workflow_dispatch:
+  pull_request:
+    branches: [main]
+    paths:
+      - ".github/workflows/xdist-experiment.yml"
+
+env:
+  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
+
+concurrency:
+  group: xdist-${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+permissions:
+  contents: read
+
+jobs:
+  xdist:
+    name: xdist (${{ matrix.os }} / py${{ matrix.python-version }})
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest, windows-latest, macos-latest]
+        python-version: ["3.12"]
+    defaults:
+      run:
+        shell: bash
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v5
+        with:
+          lfs: false
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v6
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: pip
+          cache-dependency-path: pyproject.toml
+
+      - name: Install experiment dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install ".[dev]" pytest-xdist
+
+      - name: Run xdist experiment
+        run: |
+          set +e
+          start="$(python - <<'PY'
+          import time
+          print(time.monotonic())
+          PY
+          )"
+          python -m pytest -q -m "not browser" --no-cov \
+            -n 2 \
+            --dist=loadfile \
+            --max-worker-restart=0 \
+            --durations=25
+          status="$?"
+          end="$(python - <<'PY'
+          import time
+          print(time.monotonic())
+          PY
+          )"
+          elapsed="$(python - "$start" "$end" <<'PY'
+          import sys
+          start = float(sys.argv[1])
+          end = float(sys.argv[2])
+          print(f"{end - start:.1f}")
+          PY
+          )"
+
+          {
+            echo "## pytest-xdist experiment"
+            echo ""
+            echo "- OS: $RUNNER_OS"
+            echo "- Python: ${{ matrix.python-version }}"
+            echo "- Command: \`python -m pytest -q -m \"not browser\" --no-cov -n 2 --dist=loadfile --max-worker-restart=0 --durations=25\`"
+            echo "- Exit code: $status"
+            echo "- Elapsed seconds: $elapsed"
+          } >> "$GITHUB_STEP_SUMMARY"
+
+          if [ "$status" -ne 0 ]; then
+            echo "::warning::xdist experiment failed on $RUNNER_OS with exit code $status"
+          fi
+          exit "$status"

docs/SKILL.md

@@ -1,303 +1,303 @@
----
-name: skill-router
-description: "Repo-aware recommendation manager for ctx. Scans the active repository, identifies stack and workflow signals, recommends a capped set of skills, agents, and MCP servers, and unloads helpers that no longer match the current work after user confirmation. Harnesses are recommended by the custom-model onboarding flow and then attach to the same recommendation layer."
----
-
-# Skill Router
-
-Scan a repo. Know what the current work needs. Recommend only that. Keep the
-wiki and graph as the durable catalog behind the decision.
-
-## Scope
-
-The router manages runtime recommendations for:
-
-- skills
-- agents
-- MCP servers
-
-Harnesses are separate catalog entities. They are recommended when a user wants
-to run ctx with a non-Claude-Code host, local model, or API model. Once attached,
-the harness calls the same skills/agents/MCP recommendation engine.
-
-## Problem
-
-Every loaded helper costs tokens, attention, and operational surface area. Most
-sessions need a small top-scored bundle from the shipped graph, not every entity
-ctx knows about. Loading too much:
-
-- wastes context on irrelevant instructions,
-- causes misfires when a helper matches the wrong task,
-- slows the agent loop, and
-- creates conflicting instructions.
-
-## Architecture
-
-```text
-ctx/
-|-- src/scan_repo.py                         # Repo scanner -> stack profile
-|-- src/ctx/core/resolve/resolve_skills.py   # Profile -> load/unload manifest
-|-- src/ctx/core/resolve/recommendations.py  # Shared scoring/ranking engine
-|-- src/ctx/adapters/                        # Claude Code hooks + generic tools
-|-- src/harness_install.py                   # Custom-model harness install flow
-`-- graph/wiki-graph-runtime.tar.gz          # Shipped graph/wiki runtime
-```
-
-The router has three halves:
-
-1. **Scanner** - analyzes a repo and produces stack/task evidence.
-2. **Resolver** - scores graph/wiki entities and emits a capped manifest.
-3. **Wiki/graph** - persistent catalog of entities, usage, quality, and links.
-
-## Startup Flow
-
-1. Read the shipped graph/wiki metadata and local user overrides.
-2. Run `ctx-scan-repo --repo . --recommend` for the active repo.
-3. Resolve a load/unload manifest with the shared recommendation engine.
-4. Present changes with reasons and require confirmation unless the user enabled
-   automatic mode.
-5. Record usage/quality changes after the user accepts or rejects suggestions.
-
-## Scanner
-
-The scanner reads repo structure and files to produce a stack profile. Detection
-is evidence-based: every claim should map to a file, dependency, config value, or
-import pattern.
-
-### Detection Categories
-
-- **Languages** - file extensions, shebangs, lock files.
-- **Frameworks and libraries** - package manifests, imports, config files.
-- **Infrastructure and DevOps** - Docker, CI/CD, IaC, cloud, Kubernetes.
-- **Data and storage** - databases, migrations, queues, pipelines.
-- **Documentation and content** - MkDocs, Docusaurus, Sphinx, API specs.
-- **Testing and quality** - pytest, Jest, Playwright, Ruff, mypy, TypeScript.
-- **AI and agent tooling** - MCP configs, LangGraph, CrewAI, prompt dirs, model
-  config names.
-- **Build and package** - package managers, build tools, monorepos.
-
-### Scanning Rules
-
-1. Start with filenames and config files. Read source only when needed to
-   disambiguate.
-2. Do not include speculative signals below the configured confidence floor.
-3. Skip generated/vendor directories such as `.git`, `node_modules`,
-   `__pycache__`, `venv`, and `.venv`.
-4. Keep initial scans bounded for large repos.
-5. Redact secret values; only record secret/key names as evidence.
-
-## Resolver
-
-The resolver produces a manifest containing the exact helper set to load or
-unload.
-
-```json
-{
-  "generated_at": "ISO-8601",
-  "repo_path": "/absolute/path",
-  "profile_hash": "sha256 of stack profile",
-  "load": [
-    {
-      "name": "fastapi",
-      "type": "skill",
-      "reason": "FastAPI detected in pyproject.toml dependencies",
-      "score": 0.94
-    }
-  ],
-  "unload": [
-    {
-      "name": "react",
-      "type": "skill",
-      "reason": "No active frontend signal in the current repo window"
-    }
-  ],
-  "suggestions": [
-    {
-      "name": "github",
-      "type": "mcp",
-      "reason": ".github/workflows exists and the repo uses GitHub Actions",
-      "install_command": "ctx-mcp-add ..."
-    }
-  ],
-  "warnings": []
-}
-```
-
-### Ranking Signals
-
-Candidates can be scored by:
-
-- tag/category/subcategory match
-- semantic similarity edge weight
-- direct wiki links
-- source overlap
-- type affinity
-- usage score
-- quality/security score
-- user overrides (`always_load`, `never_load`)
-- configured caps and minimum score gates
-
-Recommendations are capped. ctx should not recommend at all costs; below-threshold
-candidates should be withheld and the gap should be explained.
-
-## Wiki And Graph Contract
-
-The wiki follows a Karpathy-style durable memory pattern: schema, purpose, index,
-log, lint, review items, and entity pages that are machine-readable.
-
-```text
-llm-wiki/
-|-- SCHEMA.md
-|-- index.md
-|-- log.md
-|-- raw/
-|   |-- scans/
-|   `-- external-catalogs/
-|-- entities/
-|   |-- skills/
-|   |-- agents/
-|   |-- mcp-servers/
-|   `-- harnesses/
-|-- concepts/
-|-- comparisons/
-`-- queries/
-```
-
-Entity pages use YAML frontmatter with at least:
-
-```yaml
-title: Entity Name
-type: skill | agent | mcp-server | harness | external-catalog
-status: installed | available | deprecated | broken
-tags: []
-source: local | shipped | github | curated
-source_url: ""
-path: ""
-quality_score: null
-usage_score: 0
-last_used: null
-always_load: false
-never_load: false
-```
-
-## Entity Sources
-
-Entity sources include shipped skill/MCP/harness sources, GitHub
-entity repositories, and local user assets. The public reference page is
-[`marketplace-registry.md`](marketplace-registry.md), kept under that filename
-for backwards-compatible links.
-
-Source rules:
-
-1. Search the shipped graph/wiki first.
-2. Use additional sources only when the local graph is missing or stale.
-3. Deduplicate before adding.
-4. If an entity exists, emit an update review instead of replacing it.
-5. Run security checks before promotion.
-6. Rebuild and validate graph/wiki artifacts before shipping.
-
-## Micro-Skill Gate
-
-Every added or updated skill must pass the configured line-count gate before it
-is packed into the shipped wiki. The default threshold is 180 lines, read from
-ctx config. Skills above the threshold are converted into a short orchestrator
-plus staged reference files. Local `.original` backups may be preserved for
-traceability, but packaged runtime archives must omit those backups.
-
-## Core Operations
-
-### Full Scan
-
-Triggers: repo open, repo switch, `ctx-scan-repo --repo . --recommend`, or explicit user
-request.
-
-1. Read wiki orientation and local overrides.
-2. Scan the target repo.
-3. Save the scan result.
-4. Resolve load/unload recommendations.
-5. Present reasons, scores, and install/update commands.
-6. Apply only after confirmation unless configured otherwise.
-7. Record usage and decisions.
-
-### Incremental Scan
-
-Triggers: changed config files, new dependencies, new MCP config, new tests, new
-infrastructure files, or user task change.
-
-If a helper is no longer useful, ctx should suggest unloading it and ask for
-confirmation. If the user asks to skip unload prompts, ctx should respect that
-preference.
-
-### Manual Override
-
-- "Always load the docker skill" -> set `always_load: true`.
-- "Never load the react skill" -> set `never_load: true`.
-- "Load the terraform skill for this session" -> temporary load.
-- "What is loaded?" -> show current manifest with reasons.
-
-### Discovery
-
-When the user asks what exists for a task:
-
-1. Search graph/wiki entity pages by tags and text.
-2. Use `find-skills` and configured sources for remote freshness when needed.
-3. Show status, score, source, and risk notes.
-4. Suggest install/update commands, not silent installs.
-5. Log the query.
-
-## Maintenance Checks
-
-- **Stale installed helpers** - used rarely or not used recently.
-- **Ghost helpers** - installed status but missing local path.
-- **Orphan local helpers** - present locally but missing entity page.
-- **Skill index freshness** - shipped skill snapshot older than policy.
-- **Conflicts** - overlapping helpers both marked `always_load`.
-- **Usage cold spots** - low usage/quality score candidates for unload review.
-- **Wiki lint** - broken links, missing frontmatter, index drift.
-
-## Reporting
-
-After a scan, report only actionable details:
-
-```text
-Skill Router Report - repo-name
-Scanned: YYYY-MM-DD HH:MM
-
-Loaded:
-1. fastapi skill - score 0.94 - pyproject.toml dependency
-2. github MCP - score 0.91 - GitHub Actions workflows detected
-
-Unload candidates:
-1. react skill - no active frontend signal
-
-Suggestions:
-1. openapi-generator skill - OpenAPI spec found
-
-Warnings:
-None
-```
-
-## Configuration
-
-Router behavior is controlled by ctx config and user overrides. Important knobs:
-
-```yaml
-recommendations:
-  max_total: 5
-  min_score: 0.85
-micro_skills:
-  max_lines: 180
-router:
-  auto_scan: true
-  auto_apply: false
-  ask_before_unload: true
-```
-
-## Pitfalls
-
-- Never load all helpers just in case.
-- Never replace an existing entity without an update review.
-- Never execute external repo scripts while cataloging.
-- Never treat conversation history as the source of workflow state.
-- Never ship graph/wiki artifacts before validation passes.
-- Respect `never_load` and user rejection history.
+---
+name: skill-router
+description: "Repo-aware recommendation manager for ctx. Scans the active repository, identifies stack and workflow signals, recommends a capped set of skills, agents, and MCP servers, and unloads helpers that no longer match the current work after user confirmation. Harnesses are recommended by the custom-model onboarding flow and then attach to the same recommendation layer."
+---
+
+# Skill Router
+
+Scan a repo. Know what the current work needs. Recommend only that. Keep the
+wiki and graph as the durable catalog behind the decision.
+
+## Scope
+
+The router manages runtime recommendations for:
+
+- skills
+- agents
+- MCP servers
+
+Harnesses are separate catalog entities. They are recommended when a user wants
+to run ctx with a non-Claude-Code host, local model, or API model. Once attached,
+the harness calls the same skills/agents/MCP recommendation engine.
+
+## Problem
+
+Every loaded helper costs tokens, attention, and operational surface area. Most
+sessions need a small top-scored bundle from the shipped graph, not every entity
+ctx knows about. Loading too much:
+
+- wastes context on irrelevant instructions,
+- causes misfires when a helper matches the wrong task,
+- slows the agent loop, and
+- creates conflicting instructions.
+
+## Architecture
+
+```text
+ctx/
+|-- src/scan_repo.py                         # Repo scanner -> stack profile
+|-- src/ctx/core/resolve/resolve_skills.py   # Profile -> load/unload manifest
+|-- src/ctx/core/resolve/recommendations.py  # Shared scoring/ranking engine
+|-- src/ctx/adapters/                        # Claude Code hooks + generic tools
+|-- src/harness_install.py                   # Custom-model harness install flow
+`-- graph/wiki-graph-runtime.tar.gz          # Shipped graph/wiki runtime
+```
+
+The router has three halves:
+
+1. **Scanner** - analyzes a repo and produces stack/task evidence.
+2. **Resolver** - scores graph/wiki entities and emits a capped manifest.
+3. **Wiki/graph** - persistent catalog of entities, usage, quality, and links.
+
+## Startup Flow
+
+1. Read the shipped graph/wiki metadata and local user overrides.
+2. Run `ctx-scan-repo --repo . --recommend` for the active repo.
+3. Resolve a load/unload manifest with the shared recommendation engine.
+4. Present changes with reasons and require confirmation unless the user enabled
+   automatic mode.
+5. Record usage/quality changes after the user accepts or rejects suggestions.
+
+## Scanner
+
+The scanner reads repo structure and files to produce a stack profile. Detection
+is evidence-based: every claim should map to a file, dependency, config value, or
+import pattern.
+
+### Detection Categories
+
+- **Languages** - file extensions, shebangs, lock files.
+- **Frameworks and libraries** - package manifests, imports, config files.
+- **Infrastructure and DevOps** - Docker, CI/CD, IaC, cloud, Kubernetes.
+- **Data and storage** - databases, migrations, queues, pipelines.
+- **Documentation and content** - MkDocs, Docusaurus, Sphinx, API specs.
+- **Testing and quality** - pytest, Jest, Playwright, Ruff, mypy, TypeScript.
+- **AI and agent tooling** - MCP configs, LangGraph, CrewAI, prompt dirs, model
+  config names.
+- **Build and package** - package managers, build tools, monorepos.
+
+### Scanning Rules
+
+1. Start with filenames and config files. Read source only when needed to
+   disambiguate.
+2. Do not include speculative signals below the configured confidence floor.
+3. Skip generated/vendor directories such as `.git`, `node_modules`,
+   `__pycache__`, `venv`, and `.venv`.
+4. Keep initial scans bounded for large repos.
+5. Redact secret values; only record secret/key names as evidence.
+
+## Resolver
+
+The resolver produces a manifest containing the exact helper set to load or
+unload.
+
+```json
+{
+  "generated_at": "ISO-8601",
+  "repo_path": "/absolute/path",
+  "profile_hash": "sha256 of stack profile",
+  "load": [
+    {
+      "name": "fastapi",
+      "type": "skill",
+      "reason": "FastAPI detected in pyproject.toml dependencies",
+      "score": 0.94
+    }
+  ],
+  "unload": [
+    {
+      "name": "react",
+      "type": "skill",
+      "reason": "No active frontend signal in the current repo window"
+    }
+  ],
+  "suggestions": [
+    {
+      "name": "github",
+      "type": "mcp",
+      "reason": ".github/workflows exists and the repo uses GitHub Actions",
+      "install_command": "ctx-mcp-add ..."
+    }
+  ],
+  "warnings": []
+}
+```
+
+### Ranking Signals
+
+Candidates can be scored by:
+
+- tag/category/subcategory match
+- semantic similarity edge weight
+- direct wiki links
+- source overlap
+- type affinity
+- usage score
+- quality/security score
+- user overrides (`always_load`, `never_load`)
+- configured caps and minimum score gates
+
+Recommendations are capped. ctx should not recommend at all costs; below-threshold
+candidates should be withheld and the gap should be explained.
+
+## Wiki And Graph Contract
+
+The wiki follows a Karpathy-style durable memory pattern: schema, purpose, index,
+log, lint, review items, and entity pages that are machine-readable.
+
+```text
+llm-wiki/
+|-- SCHEMA.md
+|-- index.md
+|-- log.md
+|-- raw/
+|   |-- scans/
+|   `-- external-catalogs/
+|-- entities/
+|   |-- skills/
+|   |-- agents/
+|   |-- mcp-servers/
+|   `-- harnesses/
+|-- concepts/
+|-- comparisons/
+`-- queries/
+```
+
+Entity pages use YAML frontmatter with at least:
+
+```yaml
+title: Entity Name
+type: skill | agent | mcp-server | harness | external-catalog
+status: installed | available | deprecated | broken
+tags: []
+source: local | shipped | github | curated
+source_url: ""
+path: ""
+quality_score: null
+usage_score: 0
+last_used: null
+always_load: false
+never_load: false
+```
+
+## Entity Sources
+
+Entity sources include shipped skill/MCP/harness sources, GitHub
+entity repositories, and local user assets. The public reference page is
+[`marketplace-registry.md`](marketplace-registry.md), kept under that filename
+for backwards-compatible links.
+
+Source rules:
+
+1. Search the shipped graph/wiki first.
+2. Use additional sources only when the local graph is missing or stale.
+3. Deduplicate before adding.
+4. If an entity exists, emit an update review instead of replacing it.
+5. Run security checks before promotion.
+6. Rebuild and validate graph/wiki artifacts before shipping.
+
+## Micro-Skill Gate
+
+Every added or updated skill must pass the configured line-count gate before it
+is packed into the shipped wiki. The default threshold is 180 lines, read from
+ctx config. Skills above the threshold are converted into a short orchestrator
+plus staged reference files. Local `.original` backups may be preserved for
+traceability, but packaged runtime archives must omit those backups.
+
+## Core Operations
+
+### Full Scan
+
+Triggers: repo open, repo switch, `ctx-scan-repo --repo . --recommend`, or explicit user
+request.
+
+1. Read wiki orientation and local overrides.
+2. Scan the target repo.
+3. Save the scan result.
+4. Resolve load/unload recommendations.
+5. Present reasons, scores, and install/update commands.
+6. Apply only after confirmation unless configured otherwise.
+7. Record usage and decisions.
+
+### Incremental Scan
+
+Triggers: changed config files, new dependencies, new MCP config, new tests, new
+infrastructure files, or user task change.
+
+If a helper is no longer useful, ctx should suggest unloading it and ask for
+confirmation. If the user asks to skip unload prompts, ctx should respect that
+preference.
+
+### Manual Override
+
+- "Always load the docker skill" -> set `always_load: true`.
+- "Never load the react skill" -> set `never_load: true`.
+- "Load the terraform skill for this session" -> temporary load.
+- "What is loaded?" -> show current manifest with reasons.
+
+### Discovery
+
+When the user asks what exists for a task:
+
+1. Search graph/wiki entity pages by tags and text.
+2. Use `find-skills` and configured sources for remote freshness when needed.
+3. Show status, score, source, and risk notes.
+4. Suggest install/update commands, not silent installs.
+5. Log the query.
+
+## Maintenance Checks
+
+- **Stale installed helpers** - used rarely or not used recently.
+- **Ghost helpers** - installed status but missing local path.
+- **Orphan local helpers** - present locally but missing entity page.
+- **Skill index freshness** - shipped skill snapshot older than policy.
+- **Conflicts** - overlapping helpers both marked `always_load`.
+- **Usage cold spots** - low usage/quality score candidates for unload review.
+- **Wiki lint** - broken links, missing frontmatter, index drift.
+
+## Reporting
+
+After a scan, report only actionable details:
+
+```text
+Skill Router Report - repo-name
+Scanned: YYYY-MM-DD HH:MM
+
+Loaded:
+1. fastapi skill - score 0.94 - pyproject.toml dependency
+2. github MCP - score 0.91 - GitHub Actions workflows detected
+
+Unload candidates:
+1. react skill - no active frontend signal
+
+Suggestions:
+1. openapi-generator skill - OpenAPI spec found
+
+Warnings:
+None
+```
+
+## Configuration
+
+Router behavior is controlled by ctx config and user overrides. Important knobs:
+
+```yaml
+recommendations:
+  max_total: 5
+  min_score: 0.85
+micro_skills:
+  max_lines: 180
+router:
+  auto_scan: true
+  auto_apply: false
+  ask_before_unload: true
+```
+
+## Pitfalls
+
+- Never load all helpers just in case.
+- Never replace an existing entity without an update review.
+- Never execute external repo scripts while cataloging.
+- Never treat conversation history as the source of workflow state.
+- Never ship graph/wiki artifacts before validation passes.
+- Respect `never_load` and user rejection history.

docs/assets/javascripts/repo-stats-refresh.js

@@ -1,120 +1,120 @@
-(function () {
-  "use strict";
-
-  var REPO_URL = "https://github.com/stevesolun/ctx";
-  var REPO_API_URL = "https://api.github.com/repos/stevesolun/ctx";
-  var MATERIAL_SOURCE_CACHE_KEY = "__source";
-
-  function isCtxSource(anchor) {
-    var href = anchor.getAttribute("href") || "";
-    return href.replace(/\/$/, "") === REPO_URL;
-  }
-
-  function formatCount(value) {
-    return new Intl.NumberFormat("en", {
-      maximumFractionDigits: 1,
-      notation: value >= 10000 ? "compact" : "standard",
-    })
-      .format(value)
-      .toLowerCase();
-  }
-
-  function createFact(kind, value) {
-    var item = document.createElement("li");
-    item.className = "md-source__fact md-source__fact--" + kind;
-    item.textContent = formatCount(value);
-    return item;
-  }
-
-  function clearRenderedFacts(source) {
-    source.querySelectorAll(".md-source__facts").forEach(function (facts) {
-      facts.remove();
-    });
-
-    var repository = source.querySelector(".md-source__repository");
-    if (repository) {
-      repository.classList.remove("md-source__repository--active");
-    }
-  }
-
-  function renderStats(stats) {
-    document.querySelectorAll(".md-source").forEach(function (source) {
-      if (!(source instanceof HTMLAnchorElement) || !isCtxSource(source)) {
-        return;
-      }
-
-      clearRenderedFacts(source);
-
-      var repository = source.querySelector(".md-source__repository");
-      if (!repository) {
-        return;
-      }
-
-      var facts = document.createElement("ul");
-      facts.className = "md-source__facts";
-      facts.appendChild(createFact("stars", stats.stars));
-      facts.appendChild(createFact("forks", stats.forks));
-
-      repository.appendChild(facts);
-      repository.classList.add("md-source__repository--active");
-    });
-  }
-
-  function clearStaleStats() {
-    document.querySelectorAll(".md-source").forEach(function (source) {
-      if (source instanceof HTMLAnchorElement && isCtxSource(source)) {
-        clearRenderedFacts(source);
-      }
-    });
-  }
-
-  function updateMaterialCache(stats) {
-    try {
-      sessionStorage.setItem(
-        MATERIAL_SOURCE_CACHE_KEY,
-        JSON.stringify({ stars: stats.stars, forks: stats.forks })
-      );
-    } catch (error) {
-      /* Browser storage may be disabled. The visible header is still updated. */
-    }
-  }
-
-  async function fetchStats() {
-    var response = await fetch(REPO_API_URL, {
-      cache: "no-store",
-      headers: { Accept: "application/vnd.github+json" },
-    });
-
-    if (!response.ok) {
-      throw new Error("GitHub repository stats request failed");
-    }
-
-    var data = await response.json();
-    return {
-      forks: Number(data.forks_count || 0),
-      stars: Number(data.stargazers_count || 0),
-    };
-  }
-
-  async function refreshRepoStats() {
-    clearStaleStats();
-
-    try {
-      var stats = await fetchStats();
-      updateMaterialCache(stats);
-      renderStats(stats);
-    } catch (error) {
-      clearStaleStats();
-    }
-  }
-
-  if (document.readyState === "loading") {
-    document.addEventListener("DOMContentLoaded", refreshRepoStats, { once: true });
-  } else {
-    refreshRepoStats();
-  }
-
-  if (window.document$ && typeof window.document$.subscribe === "function") {
-    window.document$.subscribe(refreshRepoStats);
-  }
-})();
+(function () {
+  "use strict";
+
+  var REPO_URL = "https://github.com/stevesolun/ctx";
+  var REPO_API_URL = "https://api.github.com/repos/stevesolun/ctx";
+  var MATERIAL_SOURCE_CACHE_KEY = "__source";
+
+  function isCtxSource(anchor) {
+    var href = anchor.getAttribute("href") || "";
+    return href.replace(/\/$/, "") === REPO_URL;
+  }
+
+  function formatCount(value) {
+    return new Intl.NumberFormat("en", {
+      maximumFractionDigits: 1,
+      notation: value >= 10000 ? "compact" : "standard",
+    })
+      .format(value)
+      .toLowerCase();
+  }
+
+  function createFact(kind, value) {
+    var item = document.createElement("li");
+    item.className = "md-source__fact md-source__fact--" + kind;
+    item.textContent = formatCount(value);
+    return item;
+  }
+
+  function clearRenderedFacts(source) {
+    source.querySelectorAll(".md-source__facts").forEach(function (facts) {
+      facts.remove();
+    });
+
+    var repository = source.querySelector(".md-source__repository");
+    if (repository) {
+      repository.classList.remove("md-source__repository--active");
+    }
+  }
+
+  function renderStats(stats) {
+    document.querySelectorAll(".md-source").forEach(function (source) {
+      if (!(source instanceof HTMLAnchorElement) || !isCtxSource(source)) {
+        return;
+      }
+
+      clearRenderedFacts(source);
+
+      var repository = source.querySelector(".md-source__repository");
+      if (!repository) {
+        return;
+      }
+
+      var facts = document.createElement("ul");
+      facts.className = "md-source__facts";
+      facts.appendChild(createFact("stars", stats.stars));
+      facts.appendChild(createFact("forks", stats.forks));
+
+      repository.appendChild(facts);
+      repository.classList.add("md-source__repository--active");
+    });
+  }
+
+  function clearStaleStats() {
+    document.querySelectorAll(".md-source").forEach(function (source) {
+      if (source instanceof HTMLAnchorElement && isCtxSource(source)) {
+        clearRenderedFacts(source);
+      }
+    });
+  }
+
+  function updateMaterialCache(stats) {
+    try {
+      sessionStorage.setItem(
+        MATERIAL_SOURCE_CACHE_KEY,
+        JSON.stringify({ stars: stats.stars, forks: stats.forks })
+      );
+    } catch (error) {
+      /* Browser storage may be disabled. The visible header is still updated. */
+    }
+  }
+
+  async function fetchStats() {
+    var response = await fetch(REPO_API_URL, {
+      cache: "no-store",
+      headers: { Accept: "application/vnd.github+json" },
+    });
+
+    if (!response.ok) {
+      throw new Error("GitHub repository stats request failed");
+    }
+
+    var data = await response.json();
+    return {
+      forks: Number(data.forks_count || 0),
+      stars: Number(data.stargazers_count || 0),
+    };
+  }
+
+  async function refreshRepoStats() {
+    clearStaleStats();
+
+    try {
+      var stats = await fetchStats();
+      updateMaterialCache(stats);
+      renderStats(stats);
+    } catch (error) {
+      clearStaleStats();
+    }
+  }
+
+  if (document.readyState === "loading") {
+    document.addEventListener("DOMContentLoaded", refreshRepoStats, { once: true });
+  } else {
+    refreshRepoStats();
+  }
+
+  if (window.document$ && typeof window.document$.subscribe === "function") {
+    window.document$.subscribe(refreshRepoStats);
+  }
+})();

docs/backup-hook-install.md

@@ -1,224 +1,224 @@
-# Change-triggered backup — hook install
-
-One page on wiring the `backup_on_change.py` PostToolUse hook into Claude
-Code so a new snapshot fires automatically whenever you edit a tracked
-config file (`~/.claude/settings.json`, agents, skills, top-level
-manifests, etc.).
-
-## What it does
-
-On every `Edit` / `Write` / `MultiEdit` tool call, the hook:
-
-1. Reads the tool payload from stdin.
-2. Resolves `tool_input.file_path` and checks if it sits under
-   `~/.claude` in a file/tree/memory path tracked by `BackupConfig`.
-3. If tracked, shells out to
-   `python <repo>/src/backup_mirror.py snapshot-if-changed --reason <tool>:<basename>`.
-4. `snapshot-if-changed` hashes every tracked file, compares against the
-   most recent snapshot's `manifest.json`, and only creates a new folder
-   when at least one SHA differs.
-
-No-op edits don't create folders. The hook always exits 0 so a bug in
-the backup layer cannot stall a Claude session.
-
-## Register the hook
-
-Edit `~/.claude/settings.json` and add the following under `hooks` (keep
-any existing entries alongside it). Replace `<REPO>` with the absolute
-path to this checkout — on Windows this is a path like
-`C:/Steves_Files/Work/Research_and_Papers/ctx`.
-
-```json
-{
-  "hooks": {
-    "PostToolUse": [
-      {
-        "matcher": "Edit|Write|MultiEdit",
-        "hooks": [
-          {
-            "type": "command",
-            "command": "python <REPO>/hooks/backup_on_change.py"
-          }
-        ]
-      }
-    ]
-  }
-}
-```
-
-Notes:
-
-- The `matcher` is a regex against the tool name — the three names above
-  are the only tools that touch files.
-- Use forward slashes in the path even on Windows.
-- If `python` on your PATH is not the interpreter you want, give the
-  absolute path instead (e.g.
-  `C:/Users/you/.pyenv/pyenv-win/versions/3.13.2/python.exe`).
-
-## Verify it works
-
-1. Reload Claude Code (the hook registration is read at session start).
-2. Edit a tracked file, e.g. `~/.claude/CLAUDE.md`.
-3. Watch `~/.claude/backups/` — a new folder named
-   `<timestamp>__edit-claude-md` should appear within a second.
-4. Edit the same file again with identical content — no new folder
-   appears (SHA is unchanged).
-
-If nothing shows up, run the verb manually to isolate the failure:
-
-```bash
-python -m backup_mirror snapshot-if-changed --reason smoke-test --json
-```
-
-The JSON output tells you which files the detector considered new,
-changed, or removed.
-
-## What gets backed up
-
-See `src/backup_config.py` and the `backup` section of
-`src/config.json` for the current defaults:
-
-- **top_files** — `settings.json`, `skill-manifest.json`,
-  `pending-skills.json`, `CLAUDE.md`, `AGENTS.md`, `user-profile.json`,
-  `skill-system-config.json`, `skill-registry.json`.
-- **trees** — `agents/`, `skills/`.
-- **memory** — `projects/*/memory/**` when `memory_glob` is true.
-- **always excluded** — `.credentials.json`, `claude.json`, token
-  caches; these are dropped even if a user config lists them.
-
-To override per user, drop a partial config at
-`~/.claude/backup-config.json`. Fields you omit fall back to the repo
-default. Example:
-
-```json
-{
-  "retention": { "keep_latest": 100 },
-  "top_files": ["settings.json", "CLAUDE.md"]
-}
-```
-
-## Manual CLI
-
-The same verb is available as a one-shot command:
-
-```bash
-# snapshot only when something changed
-python -m backup_mirror snapshot-if-changed --reason manual-check
-
-# force an unconditional snapshot with a reason label
-python -m backup_mirror create --reason pre-upgrade
-```
-
-Both land under `~/.claude/backups/<timestamp>__<reason>/` and write a
-`manifest.json` that records the reason alongside every file's SHA-256.
-
-## Watchdog — snapshot on changes outside a Claude session
-
-The PostToolUse hook only fires on `Edit` / `Write` / `MultiEdit` tool
-calls *inside* a Claude session. If you edit `~/.claude/settings.json`
-in VS Code, or a `git pull` updates an agent file, the hook never
-sees it.
-
-For that gap, run the polling watchdog — a simple loop that calls
-`snapshot-if-changed` every N seconds:
-
-```bash
-python -m backup_mirror watchdog --interval 60
-```
-
-Flags:
-
-| Flag | Meaning |
-| --- | --- |
-| `--interval N` | Seconds between polls. Clamped to `[5, 3600]`. Default 60. |
-| `--reason-prefix LBL` | Prefix used for each snapshot's `--reason` label. Default `watchdog`. |
-| `--once` | Run exactly one tick and exit. Useful for cron / Task Scheduler. |
-| `--json` | Emit run stats as JSON on exit. |
-
-Because change detection is SHA-gated, polling is cheap — a tick with
-no real changes does zero disk writes.
-
-### Running it as a background service
-
-Ready-to-use service manifests live under
-[`docs/services/`](https://github.com/stevesolun/ctx/tree/main/docs/services).
-Each one expects you to edit a handful of paths — there's no installer
-that guesses where you keep the checkout.
-
-- **Linux (systemd user unit)** —
-  [`docs/services/systemd/claude-backup-watchdog.service`](https://github.com/stevesolun/ctx/blob/main/docs/services/systemd/claude-backup-watchdog.service).
-  Copy to `~/.config/systemd/user/`, set `CTX_REPO`, then
-  `systemctl --user enable --now claude-backup-watchdog.service`.
-- **macOS (launchd agent)** —
-  [`docs/services/macos/com.claude.backup.watchdog.plist`](https://github.com/stevesolun/ctx/blob/main/docs/services/macos/com.claude.backup.watchdog.plist).
-  Edit the `ProgramArguments` paths, drop into
-  `~/Library/LaunchAgents/`, then
-  `launchctl load -w ~/Library/LaunchAgents/com.claude.backup.watchdog.plist`.
-- **Windows (Task Scheduler installer)** —
-  [`docs/services/windows/install-backup-watchdog.ps1`](https://github.com/stevesolun/ctx/blob/main/docs/services/windows/install-backup-watchdog.ps1).
-  Run `pwsh -File docs/services/windows/install-backup-watchdog.ps1`
-  from the repo root; it detects Python on PATH, registers a
-  `ClaudeBackupWatchdog` scheduled task that runs at logon, and kicks
-  off the first tick. `-Uninstall` removes it.
-
-All three manifests assume the watchdog runs as an **unprivileged
-user** — no admin/root — because it only reads `~/.claude/` and writes
-`~/.claude/backups/`.
-
-The watchdog stops cleanly on SIGINT/SIGTERM, flushes its stats line
-to stderr, and exits 0. Pair it with the hook: the hook handles
-in-session edits in real time; the watchdog catches everything else.
-
-## Retention — how old snapshots get pruned
-
-Auto-pruning runs after every successful `snapshot-if-changed`, so the
-hook cannot fill the disk. The active policy comes from
-`BackupRetention` in `src/backup_config.py` (or your user override):
-
-| Field | Default | Meaning |
-| --- | --- | --- |
-| `keep_latest` | `50` | Always keep the N most-recent snapshots. |
-| `keep_daily` | `14` | For the M most-recent UTC days that have snapshots, keep the newest snapshot from each. |
-
-A snapshot survives the sweep iff it's in the **union** of those two
-sets. Snapshots whose `manifest.json` has a missing or zero
-`created_at` are always protected — we never silently delete something
-we can't place in time.
-
-To override per user, add a partial config at
-`~/.claude/backup-config.json`:
-
-```json
-{
-  "retention": { "keep_latest": 100, "keep_daily": 30 }
-}
-```
-
-### Manual prune
-
-```bash
-# Dry-run the configured policy — no deletions, JSON report.
-python -m backup_mirror prune --policy --dry-run --json
-
-# Apply the configured policy for real.
-python -m backup_mirror prune --policy
-
-# Legacy mode (still works): keep only the N newest.
-python -m backup_mirror prune --keep 20
-```
-
-The policy output tells you which snapshots were kept by `keep_latest`
-versus `keep_daily`, so a surprising retention decision is easy to
-audit.
-
-## Troubleshooting
-
-| Symptom | Likely cause |
-| --- | --- |
-| Hook never fires | Settings not reloaded, or `matcher` typo. |
-| Snapshot folder with no `reason` suffix | Called `create` without `--reason`. |
-| Hook fires but no folder appears | Content hash matched — nothing actually changed. |
-| Credentials appear in a snapshot | User put them in `top_files`; the `ALWAYS_EXCLUDE` filter would drop them — check you're on the current `backup_config.py`. |
-| `ImportError: backup_config` from the hook | Repo moved; update the path in `settings.json`. |
-| Snapshots pile up forever | `retention.keep_latest` / `keep_daily` too high. Run `prune --policy --dry-run --json` to see what the current policy would do, then lower the caps in `~/.claude/backup-config.json`. |
-| Prune removed too much | Run `prune --policy --dry-run` *before* committing to a new policy. A snapshot with a missing/zero `created_at` is always protected, so if it's getting deleted the manifest is probably fine and the policy is genuinely too aggressive. |
+# Change-triggered backup — hook install
+
+One page on wiring the `backup_on_change.py` PostToolUse hook into Claude
+Code so a new snapshot fires automatically whenever you edit a tracked
+config file (`~/.claude/settings.json`, agents, skills, top-level
+manifests, etc.).
+
+## What it does
+
+On every `Edit` / `Write` / `MultiEdit` tool call, the hook:
+
+1. Reads the tool payload from stdin.
+2. Resolves `tool_input.file_path` and checks if it sits under
+   `~/.claude` in a file/tree/memory path tracked by `BackupConfig`.
+3. If tracked, shells out to
+   `python <repo>/src/backup_mirror.py snapshot-if-changed --reason <tool>:<basename>`.
+4. `snapshot-if-changed` hashes every tracked file, compares against the
+   most recent snapshot's `manifest.json`, and only creates a new folder
+   when at least one SHA differs.
+
+No-op edits don't create folders. The hook always exits 0 so a bug in
+the backup layer cannot stall a Claude session.
+
+## Register the hook
+
+Edit `~/.claude/settings.json` and add the following under `hooks` (keep
+any existing entries alongside it). Replace `<REPO>` with the absolute
+path to this checkout — on Windows this is a path like
+`C:/Steves_Files/Work/Research_and_Papers/ctx`.
+
+```json
+{
+  "hooks": {
+    "PostToolUse": [
+      {
+        "matcher": "Edit|Write|MultiEdit",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "python <REPO>/hooks/backup_on_change.py"
+          }
+        ]
+      }
+    ]
+  }
+}
+```
+
+Notes:
+
+- The `matcher` is a regex against the tool name — the three names above
+  are the only tools that touch files.
+- Use forward slashes in the path even on Windows.
+- If `python` on your PATH is not the interpreter you want, give the
+  absolute path instead (e.g.
+  `C:/Users/you/.pyenv/pyenv-win/versions/3.13.2/python.exe`).
+
+## Verify it works
+
+1. Reload Claude Code (the hook registration is read at session start).
+2. Edit a tracked file, e.g. `~/.claude/CLAUDE.md`.
+3. Watch `~/.claude/backups/` — a new folder named
+   `<timestamp>__edit-claude-md` should appear within a second.
+4. Edit the same file again with identical content — no new folder
+   appears (SHA is unchanged).
+
+If nothing shows up, run the verb manually to isolate the failure:
+
+```bash
+python -m backup_mirror snapshot-if-changed --reason smoke-test --json
+```
+
+The JSON output tells you which files the detector considered new,
+changed, or removed.
+
+## What gets backed up
+
+See `src/backup_config.py` and the `backup` section of
+`src/config.json` for the current defaults:
+
+- **top_files** — `settings.json`, `skill-manifest.json`,
+  `pending-skills.json`, `CLAUDE.md`, `AGENTS.md`, `user-profile.json`,
+  `skill-system-config.json`, `skill-registry.json`.
+- **trees** — `agents/`, `skills/`.
+- **memory** — `projects/*/memory/**` when `memory_glob` is true.
+- **always excluded** — `.credentials.json`, `claude.json`, token
+  caches; these are dropped even if a user config lists them.
+
+To override per user, drop a partial config at
+`~/.claude/backup-config.json`. Fields you omit fall back to the repo
+default. Example:
+
+```json
+{
+  "retention": { "keep_latest": 100 },
+  "top_files": ["settings.json", "CLAUDE.md"]
+}
+```
+
+## Manual CLI
+
+The same verb is available as a one-shot command:
+
+```bash
+# snapshot only when something changed
+python -m backup_mirror snapshot-if-changed --reason manual-check
+
+# force an unconditional snapshot with a reason label
+python -m backup_mirror create --reason pre-upgrade
+```
+
+Both land under `~/.claude/backups/<timestamp>__<reason>/` and write a
+`manifest.json` that records the reason alongside every file's SHA-256.
+
+## Watchdog — snapshot on changes outside a Claude session
+
+The PostToolUse hook only fires on `Edit` / `Write` / `MultiEdit` tool
+calls *inside* a Claude session. If you edit `~/.claude/settings.json`
+in VS Code, or a `git pull` updates an agent file, the hook never
+sees it.
+
+For that gap, run the polling watchdog — a simple loop that calls
+`snapshot-if-changed` every N seconds:
+
+```bash
+python -m backup_mirror watchdog --interval 60
+```
+
+Flags:
+
+| Flag | Meaning |
+| --- | --- |
+| `--interval N` | Seconds between polls. Clamped to `[5, 3600]`. Default 60. |
+| `--reason-prefix LBL` | Prefix used for each snapshot's `--reason` label. Default `watchdog`. |
+| `--once` | Run exactly one tick and exit. Useful for cron / Task Scheduler. |
+| `--json` | Emit run stats as JSON on exit. |
+
+Because change detection is SHA-gated, polling is cheap — a tick with
+no real changes does zero disk writes.
+
+### Running it as a background service
+
+Ready-to-use service manifests live under
+[`docs/services/`](https://github.com/stevesolun/ctx/tree/main/docs/services).
+Each one expects you to edit a handful of paths — there's no installer
+that guesses where you keep the checkout.
+
+- **Linux (systemd user unit)** —
+  [`docs/services/systemd/claude-backup-watchdog.service`](https://github.com/stevesolun/ctx/blob/main/docs/services/systemd/claude-backup-watchdog.service).
+  Copy to `~/.config/systemd/user/`, set `CTX_REPO`, then
+  `systemctl --user enable --now claude-backup-watchdog.service`.
+- **macOS (launchd agent)** —
+  [`docs/services/macos/com.claude.backup.watchdog.plist`](https://github.com/stevesolun/ctx/blob/main/docs/services/macos/com.claude.backup.watchdog.plist).
+  Edit the `ProgramArguments` paths, drop into
+  `~/Library/LaunchAgents/`, then
+  `launchctl load -w ~/Library/LaunchAgents/com.claude.backup.watchdog.plist`.
+- **Windows (Task Scheduler installer)** —
+  [`docs/services/windows/install-backup-watchdog.ps1`](https://github.com/stevesolun/ctx/blob/main/docs/services/windows/install-backup-watchdog.ps1).
+  Run `pwsh -File docs/services/windows/install-backup-watchdog.ps1`
+  from the repo root; it detects Python on PATH, registers a
+  `ClaudeBackupWatchdog` scheduled task that runs at logon, and kicks
+  off the first tick. `-Uninstall` removes it.
+
+All three manifests assume the watchdog runs as an **unprivileged
+user** — no admin/root — because it only reads `~/.claude/` and writes
+`~/.claude/backups/`.
+
+The watchdog stops cleanly on SIGINT/SIGTERM, flushes its stats line
+to stderr, and exits 0. Pair it with the hook: the hook handles
+in-session edits in real time; the watchdog catches everything else.
+
+## Retention — how old snapshots get pruned
+
+Auto-pruning runs after every successful `snapshot-if-changed`, so the
+hook cannot fill the disk. The active policy comes from
+`BackupRetention` in `src/backup_config.py` (or your user override):
+
+| Field | Default | Meaning |
+| --- | --- | --- |
+| `keep_latest` | `50` | Always keep the N most-recent snapshots. |
+| `keep_daily` | `14` | For the M most-recent UTC days that have snapshots, keep the newest snapshot from each. |
+
+A snapshot survives the sweep iff it's in the **union** of those two
+sets. Snapshots whose `manifest.json` has a missing or zero
+`created_at` are always protected — we never silently delete something
+we can't place in time.
+
+To override per user, add a partial config at
+`~/.claude/backup-config.json`:
+
+```json
+{
+  "retention": { "keep_latest": 100, "keep_daily": 30 }
+}
+```
+
+### Manual prune
+
+```bash
+# Dry-run the configured policy — no deletions, JSON report.
+python -m backup_mirror prune --policy --dry-run --json
+
+# Apply the configured policy for real.
+python -m backup_mirror prune --policy
+
+# Legacy mode (still works): keep only the N newest.
+python -m backup_mirror prune --keep 20
+```
+
+The policy output tells you which snapshots were kept by `keep_latest`
+versus `keep_daily`, so a surprising retention decision is easy to
+audit.
+
+## Troubleshooting
+
+| Symptom | Likely cause |
+| --- | --- |
+| Hook never fires | Settings not reloaded, or `matcher` typo. |
+| Snapshot folder with no `reason` suffix | Called `create` without `--reason`. |
+| Hook fires but no folder appears | Content hash matched — nothing actually changed. |
+| Credentials appear in a snapshot | User put them in `top_files`; the `ALWAYS_EXCLUDE` filter would drop them — check you're on the current `backup_config.py`. |
+| `ImportError: backup_config` from the hook | Repo moved; update the path in `settings.json`. |
+| Snapshots pile up forever | `retention.keep_latest` / `keep_daily` too high. Run `prune --policy --dry-run --json` to see what the current policy would do, then lower the caps in `~/.claude/backup-config.json`. |
+| Prune removed too much | Run `prune --policy --dry-run` *before* committing to a new policy. A snapshot with a missing/zero `created_at` is always protected, so if it's getting deleted the manifest is probably fine and the policy is genuinely too aggressive. |

docs/dashboard.md

@@ -1,325 +1,347 @@
-# Dashboard (`ctx-monitor`)
-
-Local HTTP dashboard for ctx's currently supported live observables:
-loaded skills, agents, MCP servers, and installed harness records; session timelines; the
-knowledge graph; the LLM-wiki browser; quality grades + scores;
-durable queue state; graph/wiki artifact versions; filterable audit
-logs; generic-harness validation/escalation state; a live event stream;
-and harness wiki/graph browsing.
-
-```bash
-ctx-monitor serve              # http://127.0.0.1:8765
-ctx-monitor serve --port 8888  # custom port
-ctx-monitor serve --host 0.0.0.0 --port 8888  # LAN read-only with startup token URL
-```
-
-Zero Python dependencies added by the dashboard. Everything runs on
-stdlib `http.server`, using daemon request threads so a live
-`/api/events.stream` client cannot block normal dashboard or JSON API
-requests. The graph page uses a built-in list renderer and does not load
-third-party JavaScript.
-
-## Usage
-
-Every page in the dashboard has the same top nav, so getting around
-is `Home -> jump anywhere`. The dashboard indexes skills, agents, MCP
-servers, and harness pages in wiki/graph views. Harness installation,
-update, and uninstall run through `ctx-harness-install`; dashboard
-load/unload POSTs reject harnesses with the exact dry-run command to use.
-Quality scoring is shown for sidecar-backed skills, agents, and MCP servers.
-Generic/API/local harnesses that call ctx-core validation tools write to
-the runtime lifecycle ledger. The dashboard exposes that ledger at
-`/runtime` and as JSON at `/api/runtime.json`.
-
-### Check queue and artifact state - `/status`
-
-The status tab shows the durable wiki/graph maintenance queue and the
-generated graph/wiki artifacts that ctx can ship or consume. It reports:
-
-- queue DB availability and job counts by state (`pending`, `running`,
-  `succeeded`, `failed`, `cancelled`)
-- the 20 most recent queue jobs with kind, attempts, source, worker, and
-  last error; counts and the recent-job window are bounded in the queue DB,
-  not by loading the whole queue into dashboard memory
-- explicit crash recovery state: expired leases are requeued until their
-  retry budget is exhausted; exhausted leases become `failed`; operator
-  cancellations become terminal `cancelled` jobs
-- a visible queue DB error callout when the queue file exists but cannot be
-  opened or queried
-- artifact presence and byte size for generated
-  `~/.claude/skill-wiki/graphify-out/{graph.json,graph-delta.json,communities.json}`
-  plus the runtime skill index, falling back to the repo `graph/`
-  directory during source checkouts. The status page also reports the full
-  `wiki-graph.tar.gz` artifact when present.
-- artifact promotion metadata, including the latest promoted hash when
-  the crash-safe promotion path has recorded it
-
-### Browse the LLM wiki — `/wiki`
-
-The wiki tab requires full wiki markdown content from
-`ctx-init --graph --graph-install-mode full` or local/private wiki entities.
-The default runtime graph install powers recommendations and graph stats but
-does not expand every entity page. When entity pages exist, the wiki tab is a
-filterable card grid over a deterministic, bounded dashboard sample:
-up to 500 pages per dashboard-supported entity type under
-`~/.claude/skill-wiki/entities/{skills,agents,mcp-servers,harnesses}/`.
-MCP server pages use the sharded layout
-`entities/mcp-servers/<first-char-or-0-9>/<slug>.md`; the dashboard
-routes `/wiki/<slug>` to the same shard convention. Harness pages use
-the flat `entities/harnesses/<slug>.md` layout. Each card shows:
-
-- the slug (click to open `/wiki/<slug>?type=<entity>`)
-- the quality grade pill (A/B/C/D/F) when the entity has a sidecar,
-  otherwise a `skill`, `agent`, `mcp-server`, or `harness` type badge
-- the frontmatter `description`
-- up to 6 tags
-
-The **left sidebar** has a text search over the visible sample that
-matches slug, description, and tags, plus skill/agent/MCP/harness type
-checkboxes. Pair them to
-answer questions like "show me all grade-B agents related to
-testing" — check `agent`, type `testing` in the search box.
-
-Dashboard-supported entity pages (`/wiki/<slug>?type=<entity>`) render a
-bounded markdown preview and a bounded frontmatter table on the right, plus a
-quality banner with deep links to `/skill/<slug>` (sidecar detail) and
-`/graph?slug=<slug>&type=<entity>` (1-hop neighborhood). Long body previews and
-frontmatter values are visibly marked as truncated.
-
-### Explore the knowledge graph — `/graph`
-
-The graph tab is a built-in list view over the dashboard-supported
-skill/agent/MCP/harness graph. Imported skills are normal `skill`
-nodes in the graph. Harness nodes are browsable and filterable here;
-install/update actions remain in `ctx-harness-install`.
-When you arrive with no
-slug selected, the page shows:
-
-- a stats line with the total node + edge counts
-- a **Popular seed slugs** panel — the 18 highest-degree entities
-  rendered as clickable entity-type chips.
-  Click a chip to explore that entity's 1-hop neighborhood
-- a search box — type any valid skill, agent, MCP, or harness slug and press
-  `explore` (or hit Enter)
-- the graph list panel itself, which activates as soon as you pick a
-  seed
-
-Inside the graph list view, entity pills identify the node type. The
-focus row has `depth=0` in the page data, and neighbor rows are filterable
-by entity type and shared tag/token text.
-
-The JSON endpoint still includes blended graph edge weights, combining
-semantic similarity, explicit tag overlap, and slug-token overlap where
-available. **Tap any row** to
-navigate to that entity's wiki page. The type checkboxes hide or show
-skills, agents, MCP servers, and harnesses without reloading the graph.
-
-### Read the quality KPIs — `/kpi`
-
-The KPI tab is the browser equivalent of `python -m kpi_dashboard
-render`. It aggregates the quality + lifecycle sidecars under
-`~/.claude/skill-quality/` into a single page with six tables:
-
-1. **Header banner** — total entity count, subject breakdown, grade
-   pill counts, link to the raw `/api/kpi.json` payload, link back to
-   `/skills`.
-2. **Grade distribution** — A/B/C/D/F count and share.
-3. **Lifecycle tiers** — counts for `active`, `watch`, `demote`,
-   `archive`.
-4. **Hard floors active** — which override reasons are currently
-   pinning entities to F (`never_loaded_stale`, `intake_fail`, etc.)
-   and how many entities each one catches.
-5. **By category** — per-category count, average score, and full
-   A/B/C/D/F mix. This is the row most useful for "where are my D/F
-   skills concentrated?"
-6. **Top demotion candidates** — up to 25 active-or-watch entities
-   graded D/F, sorted by consecutive-D streak desc then raw score
-   asc. Click a slug to jump to its sidecar.
-7. **Archived** — slugs currently in the archive tier, with their
-   last-known grade.
-
-If the quality sidecar directory is empty (no scoring has happened
-yet), the page shows a helpful empty-state pointing at
-`ctx-skill-quality recompute --all`.
-
-## Routes
-
-### Top navigation
-
-Every page shows the same nav bar. The eleven tabs cover the
-dashboard-supported observable surface of ctx:
-
-```
-Home · Loaded · Skills · Wiki · Graph · Manage · Harness Setup · Docs · Config · Status · KPIs · Runtime · Sessions · Logs · Live
-```
-
-### HTML views
-
-Harness catalog entries are visible in wiki and graph routes. `/loaded` shows
-installed harness records from `~/.claude/harness-installs/*.json`, not the
-full catalog. Harness installation, update, and uninstall remain
-`ctx-harness-install` workflows, while harness scoring is not exposed in the
-dashboard yet.
-Dashboard POST actions are available only from loopback clients and require the
-per-process monitor token injected into the rendered page.
-
-| Route | What it shows |
-|---|---|
-| `/` | Home: seven stat cards (loaded, sidecars, wiki entities, graph nodes, runtime checks, audit events, sessions), grade distribution pills, recent sessions table, recent audit events |
-| `/loaded` | **Currently-loaded skills, agents, MCP servers, and installed harness records** from `~/.claude/skill-manifest.json` plus `~/.claude/harness-installs/*.json`; skill/agent/MCP rows expose supported live actions |
-| `/skills` | Every sidecar as a filterable **card grid**: left sidebar (search by slug, grade checkboxes, skill/agent/MCP toggle, hide-floored), card shows grade pill + raw score + links to sidecar/wiki/graph |
-| `/skill/<slug>` | Full sidecar breakdown: four-signal score (telemetry · intake · graph · routing), hard-floor reason, computed_at timestamp, per-skill audit timeline |
-| `/wiki` | **Wiki entity index** - bounded card-grid sample of up to 500 pages per dashboard-supported entity type under `~/.claude/skill-wiki/entities/{skills,agents,mcp-servers,harnesses}/`, including sharded MCP server pages and flat harness pages. Left sidebar: text search over the visible sample (slug, description, tag), skill/agent/MCP/harness checkboxes. |
-| `/wiki/<slug>?type=<entity>` | Dashboard-supported wiki entity page rendered: markdown body + full frontmatter table + grade banner + deep links to sidecar and graph-neighborhood views. The optional `type` query disambiguates duplicate slugs such as `langgraph`. |
-| `/graph` | **Graph explorer landing page** - node/edge count header, a "Popular seed slugs" block (18 highest-degree skill/agent/MCP/harness entities as clickable chips), search box for any skill/agent/MCP/harness slug, and the built-in graph list panel. Clicking a seed chip navigates to `/graph?slug=<slug>&type=<entity>`. |
-| `/graph?slug=<slug>&type=<entity>` | **Built-in** 1-hop neighborhood around the target skill/agent/MCP/harness slug. Entity pills identify skill, agent, MCP server, and harness rows. Tap any node to navigate to that entity's typed wiki page. Type and tag filters run client-side. |
-| `/manage` | Search, inspect, edit, delete, and manually import skill/agent/MCP/harness wiki entities through the same safe-name and mutation-token checks as live load/unload. |
-| `/harness` | Harness Setup wizard for non-Claude/custom API/local model users: collects model, goals, tool needs, safety constraints, and shows the harness recommendation/install path. |
-| `/docs` | Local repo docs rendered inside the dashboard with MkDocs-like tabs, sidebar table of contents, in-dashboard search, and source links. |
-| `/config` | Effective ctx config with defaults, required markers, field explanations, and editable user overrides where supported. |
-| `/status` | Durable queue and artifact status: job counts by state, recent queue jobs, graph/wiki artifact sizes, and crash-safe promotion metadata. |
-| `/kpi` | **KPI dashboard** — total entity count with subject breakdown, grade distribution pills, two-column tables for grade counts and lifecycle tiers (active · watch · demote · archive), hard-floor reasons with counts, **By category** table (count · avg score · A/B/C/D/F mix per category), **Top demotion candidates** (active/watch entries graded D or F, sorted by consecutive-D streak desc then score asc), and the **Archived** list. Same shape as `python -m kpi_dashboard render` but HTML |
-| `/runtime` | Generic harness runtime ledger from `CTX_RUNTIME_LIFECYCLE_DIR` or `~/.ctx/runtime/events.jsonl`: validation totals, failed/error checks, recent validation rows, and open escalations. |
-| `/sessions` | Index of every session (audit + skill-events), first/last seen, counts of skills loaded/unloaded, agents loaded/unloaded, MCPs loaded/unloaded, and lifecycle transitions |
-| `/session/<id>` | Per-session audit timeline showing the load → score_updated → unload triad with timestamps |
-| `/logs` | Last 500 audit events in a filterable table (client-side filter on event name, subject, session id) |
-| `/events` | Live SSE stream of new audit events |
-
-### JSON API
-
-| Route | Returns |
-|---|---|
-| `GET /api/sessions.json` | All sessions with aggregated counts |
-| `GET /api/manifest.json` | Raw `skill-manifest.json` passthrough |
-| `GET /api/status.json` | `{queue, artifacts}` payload: durable queue counts/recent jobs plus graph/wiki artifact file status and promotion metadata |
-| `GET /api/skill/<slug>.json` | Raw sidecar for one slug |
-| `GET /api/graph/<slug>.json?type=<entity>&hops=1&limit=40` | Dashboard-shaped skill/agent/MCP/harness `{nodes, edges, center}`; `type` is optional but recommended for duplicate slugs, `hops` is [1, 3], `limit` is [5, 150]. |
-| `GET /api/kpi.json` | `DashboardSummary` passthrough — `{total, by_subject, grade_counts, lifecycle_counts, category_breakdown, hard_floor_counts, low_quality_candidates, archived, generated_at}`. Returns `{total: 0, detail: "no sidecars yet"}` when the quality directory is empty |
-| `GET /api/runtime.json` | Runtime lifecycle summary: source path, validation count, failed/error count, open-escalation count, latest validation, recent validations, open escalations, and session IDs. |
-| `GET /api/config.json` | Effective/default/user config payload used by the Config tab. |
-| `GET /api/entities/search.json?q=<text>&type=<entity>&limit=80` | Wiki entity search results for Manage, Config, and entity picker flows. |
-| `GET /api/entity/<slug>.json?type=<entity>` | Frontmatter and Markdown body for one wiki entity. |
-| `GET /api/events.stream` | Server-sent events tail of `~/.claude/ctx-audit.jsonl` |
-
-### Mutation endpoints
-
-Dashboard GET views are read-only. When `ctx-monitor` is bound to a
-non-loopback host, HTML, `/api/*` JSON, and SSE routes require the
-read-token URL printed by `ctx-monitor`; the first successful token URL
-sets an HttpOnly same-site cookie for dashboard navigation. Keep the
-default loopback bind for local automation. POST endpoints enforce
-same-origin (browser tab open on another origin can't forge a request), require the per-process
-`X-CTX-Monitor-Token` injected into the dashboard page, and reject any
-slug failing the shared safe-name validator. That validator blocks path
-separators, Windows drive-relative strings, malformed names, and Windows
-reserved device names such as `con.txt` and `nul.`. There is no harness
-load/unload mutation endpoint yet.
-
-| Route | Body | Calls |
-|---|---|---|
-| `POST /api/load` | `{"slug": "...", "entity_type": "skill"}` | `skill_install.install_skill(slug)` |
-| `POST /api/load` | `{"slug": "...", "entity_type": "agent"}` | `agent_install.install_agent(slug)` |
-| `POST /api/load` | `{"slug": "...", "entity_type": "mcp-server"}` | `mcp_install.install_mcp(slug, command?, json_config?, auto=True)` |
-| `POST /api/unload` | `{"slug": "...", "entity_type": "skill"}` | `skill_unload.unload_from_session([slug])` |
-| `POST /api/unload` | `{"slug": "...", "entity_type": "agent"}` | remove the agent row from `skill-manifest.json` and append an unload row |
-| `POST /api/unload` | `{"slug": "...", "entity_type": "mcp-server"}` | `mcp_install.uninstall_mcp(slug, wiki_dir=...)` |
-| `POST /api/config` | `{"updates": {...}}` | persist supported user config overrides after validation |
-| `POST /api/entity/upsert` | entity metadata/body payload | write or update a wiki entity, then attach graph/recommendation metadata |
-| `POST /api/entity/delete` | `{"slug": "...", "entity_type": "skill"}` | remove a dashboard-supported wiki entity after safe-name validation |
-
-Harness load/unload POSTs are rejected with the exact
-`ctx-harness-install ... --dry-run` command to run instead. Skill rows emit
-`skill.loaded` / `skill.unloaded`, agent rows emit `agent.loaded` /
-`agent.unloaded`, and MCP rows emit `toolbox.triggered` with
-`meta.entity_type="mcp-server"` and `meta.action` set to `loaded` or
-`unloaded`. All dashboard-driven rows use `actor=user` and
-`meta.via="ctx-monitor"` so they appear in the session timeline.
-
-## KPIs, measures, scores
-
-The dashboard surfaces every quality signal ctx currently computes for
-sidecar-backed skills, agents, and MCP servers. Harness scoring is not
-yet exposed in the dashboard. Nothing is aggregated-only — you can
-always drill from a headline number to the raw sidecar that produced it.
-
-### On the home page
-
-| Card | What it means |
-|---|---|
-| **Currently loaded** | Count of entries in `skill-manifest.json[load]`. Clicking the card drills to `/loaded` |
-| **Sidecars** | Total sidecars in `~/.claude/skill-quality/` |
-| **Wiki entities** | Count of dashboard-supported wiki pages (skills + agents + MCP servers + harnesses) |
-| **Knowledge graph** | Dashboard-supported skill/agent/MCP/harness node count + edge count from `graphify-out/graph.json` |
-| **Runtime checks** | Validation totals, failed/error checks, and open escalations from the generic runtime lifecycle ledger |
-| **Audit events** | Line count of `~/.claude/ctx-audit.jsonl` |
-| **Sessions** | Unique session IDs seen across audit + events |
-| **Grade pills** | A / B / C / D / F counts across all sidecars, colored |
-
-### On `/skills`
-
-Every card shows:
-
-- **grade** — A / B / C / D / F pill (A=green, F=red)
-- **raw score** — float in [0, 1] before the hard-floor override
-- **subject_type** — skill, agent, or mcp-server
-- **hard floor reason** — `never_loaded_stale`, `intake_fail`, etc.
-  when the floor is active
-
-Cards sorted by `(grade, -raw_score)` so high-scoring A's come first.
-
-### On `/skill/<slug>`
-
-The full four-signal breakdown from the sidecar:
-
-| Signal | Weight (default) | What it measures |
-|---|---:|---|
-| **Telemetry** | 0.40 | Load frequency + recency from `skill-events.jsonl`. Rewards skills that are actually used. |
-| **Intake** | 0.20 | Structural health: frontmatter fields present, H1 present, minimum body length, description length. Zero if `intake_fail` floor is active. |
-| **Graph** | 0.25 | Connectivity in the knowledge graph: degree, average edge weight, community size |
-| **Routing** | 0.15 | Router hit rate from `~/.claude/router-trace.jsonl`: how often this skill was among the top-K recommendations when surfaced |
-
-The final score is `sum(weight[i] * signal[i])`. A hard floor
-(`never_loaded_stale`, `intake_fail`) can override the score to
-force an F grade regardless of other signals.
-
-The skill detail page also shows the audit timeline for this slug
-specifically: every `skill.loaded`, `skill.unloaded`,
-`skill.score_updated` row with its session_id, so you can trace
-exactly why the score changed when it did.
-
-### On `/session/<id>`
-
-The per-session view lets you watch a skill's lifecycle inside one
-session:
-
-```
-skill.loaded        fastapi-pro       session-abc  @ 10:23:05
-skill.score_updated fastapi-pro       session-abc  @ 10:31:47   grade C->B
-skill.unloaded      fastapi-pro       session-abc  @ 11:04:02
-```
-
-The `load → score_updated → unload` triad is the canonical
-observability proof that ctx's telemetry pipeline is live.
-
-## Security
-
-- **Binds to 127.0.0.1 by default**. Use `--host 0.0.0.0` only if
-  you actually want LAN-visible read-only access. The startup output
-  prints a one-process read-token URL; without that token or the cookie
-  it sets, LAN HTML/API/SSE requests return 403. Mutations remain
-  disabled on non-loopback binds.
-- **Same-origin gating on mutation**. Any POST with an `Origin`
-  header that doesn't match `Host` returns 403. Curl and direct
-  tool calls are allowed (no Origin header at all).
-- **Slug allowlist on all paths**. Anywhere the dashboard resolves
-  a skill, agent, MCP, or harness slug to a file path (`/wiki/<slug>`,
-  `/graph?slug=<slug>&type=<entity>`, `/api/graph/<slug>.json`), the slug is
-  validated through the shared
-  safe-name helper — no path traversal, no absolute paths, no UNC
-  shares, no Windows reserved device names.
-
-## Stopping
-
-Ctrl+C in the terminal. Request handling is threaded for local dashboard
-responsiveness, and shutdown signals any open SSE workers. The monitor is
-still not suitable for shared/production serving.
+# Dashboard (`ctx-monitor`)
+
+Local HTTP dashboard for ctx's currently supported live observables:
+loaded skills, agents, MCP servers, and installed harness records; session timelines; the
+knowledge graph; the LLM-wiki browser; quality grades + scores;
+durable queue state; graph/wiki artifact versions; filterable audit
+logs; generic-harness validation/escalation state; a live event stream;
+and harness wiki/graph browsing.
+
+```bash
+ctx-monitor serve              # http://127.0.0.1:8765
+ctx-monitor serve --port 8888  # custom port
+ctx-monitor serve --host 0.0.0.0 --port 8888  # LAN read-only with startup token URL
+```
+
+Zero Python dependencies added by the dashboard. Everything runs on
+stdlib `http.server`, using daemon request threads so a live
+`/api/events.stream` client cannot block normal dashboard or JSON API
+requests. The graph page uses a built-in list renderer and does not load
+third-party JavaScript.
+
+## Usage
+
+Every page in the dashboard has the same top nav, so getting around
+is `Home -> jump anywhere`. The dashboard indexes skills, agents, MCP
+servers, and harness pages in wiki/graph views. Harness installation,
+update, and uninstall run through `ctx-harness-install`; dashboard
+load/unload POSTs reject harnesses with the exact dry-run command to use.
+Quality scoring is shown for sidecar-backed skills, agents, and MCP servers.
+Generic/API/local harnesses that call ctx-core validation tools write to
+the runtime lifecycle ledger. The dashboard exposes that ledger at
+`/runtime` and as JSON at `/api/runtime.json`.
+
+### Check queue and artifact state - `/status`
+
+The status tab shows the durable wiki/graph maintenance queue and the
+generated graph/wiki artifacts that ctx can ship or consume. It reports:
+
+- queue DB availability and job counts by state (`pending`, `running`,
+  `succeeded`, `failed`, `cancelled`)
+- the 20 most recent queue jobs with kind, attempts, source, worker, and
+  last error; counts and the recent-job window are bounded in the queue DB,
+  not by loading the whole queue into dashboard memory
+- explicit crash recovery state: expired leases are requeued until their
+  retry budget is exhausted; exhausted leases become `failed`; operator
+  cancellations become terminal `cancelled` jobs
+- a visible queue DB error callout when the queue file exists but cannot be
+  opened or queried
+- artifact presence and byte size for generated
+  `~/.claude/skill-wiki/graphify-out/{graph.json,graph-delta.json,communities.json}`
+  plus the runtime skill index, falling back to the repo `graph/`
+  directory during source checkouts. The status page also reports the full
+  `wiki-graph.tar.gz` artifact when present.
+- artifact promotion metadata, including the latest promoted hash when
+  the crash-safe promotion path has recorded it
+
+### Browse the LLM wiki — `/wiki`
+
+#### Catalog badge links
+
+The README entity badges open this public docs section so they never point at a
+dead `127.0.0.1` URL from GitHub, PyPI, or Hugging Face. To open the live
+searchable tile catalog, install the full wiki pages and start the local
+dashboard:
+
+```bash
+ctx-init --graph --graph-install-mode full --model-mode skip
+ctx-monitor serve
+```
+
+Then use:
+
+- Skills: `http://127.0.0.1:8765/wiki?type=skill`
+- Agents: `http://127.0.0.1:8765/wiki?type=agent`
+- MCP servers: `http://127.0.0.1:8765/wiki?type=mcp-server`
+- Harnesses: `http://127.0.0.1:8765/wiki?type=harness`
+
+The local catalog includes search, browser autocomplete suggestions, type
+filters, tile cards, and click-through detail pages for each entity.
+
+The wiki tab requires full wiki markdown content from
+`ctx-init --graph --graph-install-mode full` or local/private wiki entities.
+The default runtime graph install powers recommendations and graph stats but
+does not expand every entity page. When entity pages exist, the wiki tab is a
+filterable card grid over a deterministic, bounded dashboard sample:
+up to 500 pages per dashboard-supported entity type under
+`~/.claude/skill-wiki/entities/{skills,agents,mcp-servers,harnesses}/`.
+MCP server pages use the sharded layout
+`entities/mcp-servers/<first-char-or-0-9>/<slug>.md`; the dashboard
+routes `/wiki/<slug>` to the same shard convention. Harness pages use
+the flat `entities/harnesses/<slug>.md` layout. Each card shows:
+
+- the slug (click to open `/wiki/<slug>?type=<entity>`)
+- the quality grade pill (A/B/C/D/F) when the entity has a sidecar,
+  otherwise a `skill`, `agent`, `mcp-server`, or `harness` type badge
+- the frontmatter `description`
+- up to 6 tags
+
+The **left sidebar** has a text search over the visible sample that
+matches slug, description, and tags, plus skill/agent/MCP/harness type
+checkboxes. Pair them to
+answer questions like "show me all grade-B agents related to
+testing" — check `agent`, type `testing` in the search box.
+
+Dashboard-supported entity pages (`/wiki/<slug>?type=<entity>`) render a
+bounded markdown preview and a bounded frontmatter table on the right, plus a
+quality banner with deep links to `/skill/<slug>` (sidecar detail) and
+`/graph?slug=<slug>&type=<entity>` (1-hop neighborhood). Long body previews and
+frontmatter values are visibly marked as truncated.
+
+### Explore the knowledge graph — `/graph`
+
+The graph tab is a built-in list view over the dashboard-supported
+skill/agent/MCP/harness graph. Imported skills are normal `skill`
+nodes in the graph. Harness nodes are browsable and filterable here;
+install/update actions remain in `ctx-harness-install`.
+When you arrive with no
+slug selected, the page shows:
+
+- a stats line with the total node + edge counts
+- a **Popular seed slugs** panel — the 18 highest-degree entities
+  rendered as clickable entity-type chips.
+  Click a chip to explore that entity's 1-hop neighborhood
+- a search box — type any valid skill, agent, MCP, or harness slug and press
+  `explore` (or hit Enter)
+- the graph list panel itself, which activates as soon as you pick a
+  seed
+
+Inside the graph list view, entity pills identify the node type. The
+focus row has `depth=0` in the page data, and neighbor rows are filterable
+by entity type and shared tag/token text.
+
+The JSON endpoint still includes blended graph edge weights, combining
+semantic similarity, explicit tag overlap, and slug-token overlap where
+available. **Tap any row** to
+navigate to that entity's wiki page. The type checkboxes hide or show
+skills, agents, MCP servers, and harnesses without reloading the graph.
+
+### Read the quality KPIs — `/kpi`
+
+The KPI tab is the browser equivalent of `python -m kpi_dashboard
+render`. It aggregates the quality + lifecycle sidecars under
+`~/.claude/skill-quality/` into a single page with six tables:
+
+1. **Header banner** — total entity count, subject breakdown, grade
+   pill counts, link to the raw `/api/kpi.json` payload, link back to
+   `/skills`.
+2. **Grade distribution** — A/B/C/D/F count and share.
+3. **Lifecycle tiers** — counts for `active`, `watch`, `demote`,
+   `archive`.
+4. **Hard floors active** — which override reasons are currently
+   pinning entities to F (`never_loaded_stale`, `intake_fail`, etc.)
+   and how many entities each one catches.
+5. **By category** — per-category count, average score, and full
+   A/B/C/D/F mix. This is the row most useful for "where are my D/F
+   skills concentrated?"
+6. **Top demotion candidates** — up to 25 active-or-watch entities
+   graded D/F, sorted by consecutive-D streak desc then raw score
+   asc. Click a slug to jump to its sidecar.
+7. **Archived** — slugs currently in the archive tier, with their
+   last-known grade.
+
+If the quality sidecar directory is empty (no scoring has happened
+yet), the page shows a helpful empty-state pointing at
+`ctx-skill-quality recompute --all`.
+
+## Routes
+
+### Top navigation
+
+Every page shows the same nav bar. The eleven tabs cover the
+dashboard-supported observable surface of ctx:
+
+```
+Home · Loaded · Skills · Wiki · Graph · Manage · Harness Setup · Docs · Config · Status · KPIs · Runtime · Sessions · Logs · Live
+```
+
+### HTML views
+
+Harness catalog entries are visible in wiki and graph routes. `/loaded` shows
+installed harness records from `~/.claude/harness-installs/*.json`, not the
+full catalog. Harness installation, update, and uninstall remain
+`ctx-harness-install` workflows, while harness scoring is not exposed in the
+dashboard yet.
+Dashboard POST actions are available only from loopback clients and require the
+per-process monitor token injected into the rendered page.
+
+| Route | What it shows |
+|---|---|
+| `/` | Home: seven stat cards (loaded, sidecars, wiki entities, graph nodes, runtime checks, audit events, sessions), grade distribution pills, recent sessions table, recent audit events |
+| `/loaded` | **Currently-loaded skills, agents, MCP servers, and installed harness records** from `~/.claude/skill-manifest.json` plus `~/.claude/harness-installs/*.json`; skill/agent/MCP rows expose supported live actions |
+| `/skills` | Every sidecar as a filterable **card grid**: left sidebar (search by slug, grade checkboxes, skill/agent/MCP toggle, hide-floored), card shows grade pill + raw score + links to sidecar/wiki/graph |
+| `/skill/<slug>` | Full sidecar breakdown: four-signal score (telemetry · intake · graph · routing), hard-floor reason, computed_at timestamp, per-skill audit timeline |
+| `/wiki` | **Wiki entity index** - bounded card-grid sample of up to 500 pages per dashboard-supported entity type under `~/.claude/skill-wiki/entities/{skills,agents,mcp-servers,harnesses}/`, including sharded MCP server pages and flat harness pages. Left sidebar: text search over the visible sample (slug, description, tag), skill/agent/MCP/harness checkboxes. |
+| `/wiki/<slug>?type=<entity>` | Dashboard-supported wiki entity page rendered: markdown body + full frontmatter table + grade banner + deep links to sidecar and graph-neighborhood views. The optional `type` query disambiguates duplicate slugs such as `langgraph`. |
+| `/graph` | **Graph explorer landing page** - node/edge count header, a "Popular seed slugs" block (18 highest-degree skill/agent/MCP/harness entities as clickable chips), search box for any skill/agent/MCP/harness slug, and the built-in graph list panel. Clicking a seed chip navigates to `/graph?slug=<slug>&type=<entity>`. |
+| `/graph?slug=<slug>&type=<entity>` | **Built-in** 1-hop neighborhood around the target skill/agent/MCP/harness slug. Entity pills identify skill, agent, MCP server, and harness rows. Tap any node to navigate to that entity's typed wiki page. Type and tag filters run client-side. |
+| `/manage` | Search, inspect, edit, delete, and manually import skill/agent/MCP/harness wiki entities through the same safe-name and mutation-token checks as live load/unload. |
+| `/harness` | Harness Setup wizard for non-Claude/custom API/local model users: collects model, goals, tool needs, safety constraints, and shows the harness recommendation/install path. |
+| `/docs` | Local repo docs rendered inside the dashboard with MkDocs-like tabs, sidebar table of contents, in-dashboard search, and source links. |
+| `/config` | Effective ctx config with defaults, required markers, field explanations, and editable user overrides where supported. |
+| `/status` | Durable queue and artifact status: job counts by state, recent queue jobs, graph/wiki artifact sizes, and crash-safe promotion metadata. |
+| `/kpi` | **KPI dashboard** — total entity count with subject breakdown, grade distribution pills, two-column tables for grade counts and lifecycle tiers (active · watch · demote · archive), hard-floor reasons with counts, **By category** table (count · avg score · A/B/C/D/F mix per category), **Top demotion candidates** (active/watch entries graded D or F, sorted by consecutive-D streak desc then score asc), and the **Archived** list. Same shape as `python -m kpi_dashboard render` but HTML |
+| `/runtime` | Generic harness runtime ledger from `CTX_RUNTIME_LIFECYCLE_DIR` or `~/.ctx/runtime/events.jsonl`: validation totals, failed/error checks, recent validation rows, and open escalations. |
+| `/sessions` | Index of every session (audit + skill-events), first/last seen, counts of skills loaded/unloaded, agents loaded/unloaded, MCPs loaded/unloaded, and lifecycle transitions |
+| `/session/<id>` | Per-session audit timeline showing the load → score_updated → unload triad with timestamps |
+| `/logs` | Last 500 audit events in a filterable table (client-side filter on event name, subject, session id) |
+| `/events` | Live SSE stream of new audit events |
+
+### JSON API
+
+| Route | Returns |
+|---|---|
+| `GET /api/sessions.json` | All sessions with aggregated counts |
+| `GET /api/manifest.json` | Raw `skill-manifest.json` passthrough |
+| `GET /api/status.json` | `{queue, artifacts}` payload: durable queue counts/recent jobs plus graph/wiki artifact file status and promotion metadata |
+| `GET /api/skill/<slug>.json` | Raw sidecar for one slug |
+| `GET /api/graph/<slug>.json?type=<entity>&hops=1&limit=40` | Dashboard-shaped skill/agent/MCP/harness `{nodes, edges, center}`; `type` is optional but recommended for duplicate slugs, `hops` is [1, 3], `limit` is [5, 150]. |
+| `GET /api/kpi.json` | `DashboardSummary` passthrough — `{total, by_subject, grade_counts, lifecycle_counts, category_breakdown, hard_floor_counts, low_quality_candidates, archived, generated_at}`. Returns `{total: 0, detail: "no sidecars yet"}` when the quality directory is empty |
+| `GET /api/runtime.json` | Runtime lifecycle summary: source path, validation count, failed/error count, open-escalation count, latest validation, recent validations, open escalations, and session IDs. |
+| `GET /api/config.json` | Effective/default/user config payload used by the Config tab. |
+| `GET /api/entities/search.json?q=<text>&type=<entity>&limit=80` | Wiki entity search results for Manage, Config, and entity picker flows. |
+| `GET /api/entity/<slug>.json?type=<entity>` | Frontmatter and Markdown body for one wiki entity. |
+| `GET /api/events.stream` | Server-sent events tail of `~/.claude/ctx-audit.jsonl` |
+
+### Mutation endpoints
+
+Dashboard GET views are read-only. When `ctx-monitor` is bound to a
+non-loopback host, HTML, `/api/*` JSON, and SSE routes require the
+read-token URL printed by `ctx-monitor`; the first successful token URL
+sets an HttpOnly same-site cookie for dashboard navigation. Keep the
+default loopback bind for local automation. POST endpoints enforce
+same-origin (browser tab open on another origin can't forge a request), require the per-process
+`X-CTX-Monitor-Token` injected into the dashboard page, and reject any
+slug failing the shared safe-name validator. That validator blocks path
+separators, Windows drive-relative strings, malformed names, and Windows
+reserved device names such as `con.txt` and `nul.`. There is no harness
+load/unload mutation endpoint yet.
+
+| Route | Body | Calls |
+|---|---|---|
+| `POST /api/load` | `{"slug": "...", "entity_type": "skill"}` | `skill_install.install_skill(slug)` |
+| `POST /api/load` | `{"slug": "...", "entity_type": "agent"}` | `agent_install.install_agent(slug)` |
+| `POST /api/load` | `{"slug": "...", "entity_type": "mcp-server"}` | `mcp_install.install_mcp(slug, command?, json_config?, auto=True)` |
+| `POST /api/unload` | `{"slug": "...", "entity_type": "skill"}` | `skill_unload.unload_from_session([slug])` |
+| `POST /api/unload` | `{"slug": "...", "entity_type": "agent"}` | remove the agent row from `skill-manifest.json` and append an unload row |
+| `POST /api/unload` | `{"slug": "...", "entity_type": "mcp-server"}` | `mcp_install.uninstall_mcp(slug, wiki_dir=...)` |
+| `POST /api/config` | `{"updates": {...}}` | persist supported user config overrides after validation |
+| `POST /api/entity/upsert` | entity metadata/body payload | write or update a wiki entity, then attach graph/recommendation metadata |
+| `POST /api/entity/delete` | `{"slug": "...", "entity_type": "skill"}` | remove a dashboard-supported wiki entity after safe-name validation |
+
+Harness load/unload POSTs are rejected with the exact
+`ctx-harness-install ... --dry-run` command to run instead. Skill rows emit
+`skill.loaded` / `skill.unloaded`, agent rows emit `agent.loaded` /
+`agent.unloaded`, and MCP rows emit `toolbox.triggered` with
+`meta.entity_type="mcp-server"` and `meta.action` set to `loaded` or
+`unloaded`. All dashboard-driven rows use `actor=user` and
+`meta.via="ctx-monitor"` so they appear in the session timeline.
+
+## KPIs, measures, scores
+
+The dashboard surfaces every quality signal ctx currently computes for
+sidecar-backed skills, agents, and MCP servers. Harness scoring is not
+yet exposed in the dashboard. Nothing is aggregated-only — you can
+always drill from a headline number to the raw sidecar that produced it.
+
+### On the home page
+
+| Card | What it means |
+|---|---|
+| **Currently loaded** | Count of entries in `skill-manifest.json[load]`. Clicking the card drills to `/loaded` |
+| **Sidecars** | Total sidecars in `~/.claude/skill-quality/` |
+| **Wiki entities** | Count of dashboard-supported wiki pages (skills + agents + MCP servers + harnesses) |
+| **Knowledge graph** | Dashboard-supported skill/agent/MCP/harness node count + edge count from `graphify-out/graph.json` |
+| **Runtime checks** | Validation totals, failed/error checks, and open escalations from the generic runtime lifecycle ledger |
+| **Audit events** | Line count of `~/.claude/ctx-audit.jsonl` |
+| **Sessions** | Unique session IDs seen across audit + events |
+| **Grade pills** | A / B / C / D / F counts across all sidecars, colored |
+
+### On `/skills`
+
+Every card shows:
+
+- **grade** — A / B / C / D / F pill (A=green, F=red)
+- **raw score** — float in [0, 1] before the hard-floor override
+- **subject_type** — skill, agent, or mcp-server
+- **hard floor reason** — `never_loaded_stale`, `intake_fail`, etc.
+  when the floor is active
+
+Cards sorted by `(grade, -raw_score)` so high-scoring A's come first.
+
+### On `/skill/<slug>`
+
+The full four-signal breakdown from the sidecar:
+
+| Signal | Weight (default) | What it measures |
+|---|---:|---|
+| **Telemetry** | 0.40 | Load frequency + recency from `skill-events.jsonl`. Rewards skills that are actually used. |
+| **Intake** | 0.20 | Structural health: frontmatter fields present, H1 present, minimum body length, description length. Zero if `intake_fail` floor is active. |
+| **Graph** | 0.25 | Connectivity in the knowledge graph: degree, average edge weight, community size |
+| **Routing** | 0.15 | Router hit rate from `~/.claude/router-trace.jsonl`: how often this skill was among the top-K recommendations when surfaced |
+
+The final score is `sum(weight[i] * signal[i])`. A hard floor
+(`never_loaded_stale`, `intake_fail`) can override the score to
+force an F grade regardless of other signals.
+
+The skill detail page also shows the audit timeline for this slug
+specifically: every `skill.loaded`, `skill.unloaded`,
+`skill.score_updated` row with its session_id, so you can trace
+exactly why the score changed when it did.
+
+### On `/session/<id>`
+
+The per-session view lets you watch a skill's lifecycle inside one
+session:
+
+```
+skill.loaded        fastapi-pro       session-abc  @ 10:23:05
+skill.score_updated fastapi-pro       session-abc  @ 10:31:47   grade C->B
+skill.unloaded      fastapi-pro       session-abc  @ 11:04:02
+```
+
+The `load → score_updated → unload` triad is the canonical
+observability proof that ctx's telemetry pipeline is live.
+
+## Security
+
+- **Binds to 127.0.0.1 by default**. Use `--host 0.0.0.0` only if
+  you actually want LAN-visible read-only access. The startup output
+  prints a one-process read-token URL; without that token or the cookie
+  it sets, LAN HTML/API/SSE requests return 403. Mutations remain
+  disabled on non-loopback binds.
+- **Same-origin gating on mutation**. Any POST with an `Origin`
+  header that doesn't match `Host` returns 403. Curl and direct
+  tool calls are allowed (no Origin header at all).
+- **Slug allowlist on all paths**. Anywhere the dashboard resolves
+  a skill, agent, MCP, or harness slug to a file path (`/wiki/<slug>`,
+  `/graph?slug=<slug>&type=<entity>`, `/api/graph/<slug>.json`), the slug is
+  validated through the shared
+  safe-name helper — no path traversal, no absolute paths, no UNC
+  shares, no Windows reserved device names.
+
+## Stopping
+
+Ctrl+C in the terminal. Request handling is threaded for local dashboard
+responsiveness, and shutdown signals any open SSE workers. The monitor is
+still not suitable for shared/production serving.

docs/entity-onboarding.md

@@ -1,394 +1,394 @@
-# Entity Onboarding
-
-ctx treats skills, agents, MCP servers, and harnesses as wiki entities that can
-be indexed, linked in the knowledge graph, and recommended from the same
-surface. The important distinction is install behavior:
-
-- Skills and agents are local Claude Code assets.
-- MCP servers are recorded first, then installed only when the user opts in.
-- Harnesses are recorded first. A harness describes the machinery around the
-  model: runtime, tools, access boundaries, memory, verification, and approval
-  policy. Adding one never executes upstream setup commands.
-
-After adding any entity, drain the durable wiki queue when you want the local
-runtime graph to see it immediately:
-
-```bash
-ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1
-ctx-scan-repo --repo . --recommend
-```
-
-If a persisted semantic vector index exists, that worker pass also runs a
-best-effort incremental attach into `graphify-out/entity-overlays.jsonl` so the
-new entity can connect to existing graph nodes without a full all-pairs
-semantic rebuild. The wiki page remains the source of truth; if incremental
-attach is skipped or fails, the worker still queues the normal incremental
-graph export job.
-
-## Updating the Graph and LLM Wiki
-
-Use this sequence for every accepted skill, agent, MCP server, or harness
-change. The graph and LLM-wiki are shippable artifacts, not scratch output, so
-the update is treated like a release step.
-
-1. Add or update the entity through the matching command:
-   `ctx-skill-add`, `ctx-agent-add`, `ctx-mcp-add`, or `ctx-harness-add`.
-2. If the entity already exists, read the update review. It lists changed
-   fields, likely benefits, regressions, and security findings. Do not pass
-   `--update-existing` until those findings are acceptable.
-3. Run the security/cyber check below.
-4. Park heavyweight graph artifacts locally before rebuilds:
-   `python scripts/graph_artifact_guard.py park`. This keeps background Git
-   integrations from repeatedly LFS-cleaning the full wiki tarball while it is
-   still changing.
-5. Drain the wiki queue for local runtime use:
-   `ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1`. This updates the
-   wiki index, attempts incremental ANN graph attach when a vector index exists,
-   and queues the normal incremental graph export.
-6. Rebuild the curated wiki graph with `ctx-wiki-graphify` before shipping
-   release artifacts or when you need a full graph/export reconciliation.
-7. Repack `graph/wiki-graph.tar.gz` through the artifact promotion path:
-   write a staged tarball, validate it, atomically promote it, and keep the
-   generated `*.promotion.json` metadata with the previous/current hashes.
-   Never commit local review reports or raw caches.
-8. Refresh the bulk skill index when shipping large skill updates.
-   This adds first-class `skill` nodes, skill pages under `entities/skills/`, install
-   commands, duplicate hints, and metadata-only quality/security signals:
-
-   ```bash
-   python src/import_skills_sh_catalog.py --from-api-union <raw.json> \
-     --catalog-out graph/<skill-index>.json.gz \
-     --wiki-tar graph/wiki-graph.tar.gz \
-     --update-wiki-tar
-   ```
-9. Refresh published counts with `python src/update_repo_stats.py`.
-10. Verify the changed entity can be recommended through
-   `ctx-scan-repo --repo . --recommend` or `ctx__recommend_bundle`.
-11. Unpark and stage the graph artifacts once the release candidate is final:
-    `python scripts/graph_artifact_guard.py unpark`, then `git add` the graph
-    artifacts intentionally. Run `python scripts/graph_artifact_guard.py prune`
-    after interrupted Git/LFS runs or after release staging to clean prunable
-    local LFS cache entries. Add `--include-git-prune` only when you explicitly
-    want repo-wide dangling Git objects removed too.
-
-The durable wiki worker drains `entity-upsert`, `graph-export`,
-`skill-index-refresh`, `tar-refresh`, and `artifact-promotion` jobs. Use
-`ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1` for a controlled
-single-job drain, or omit `--limit` to drain the ready queue.
-
-For a manual attach dry-run against an existing vector index:
-
-```bash
-ctx-incremental-attach attach \
-  --index-dir ~/.claude/skill-wiki/.embedding-cache/graph/vector-index \
-  --overlay ~/.claude/skill-wiki/graphify-out/entity-overlays.jsonl \
-  --node-id skill:fastapi-review \
-  --type skill \
-  --label fastapi-review \
-  --text-file ~/.claude/skill-wiki/entities/skills/fastapi-review.md \
-  --dry-run
-```
-
-Use `ctx-incremental-attach calibrate --graph ~/.claude/skill-wiki/graphify-out/graph.json`
-to inspect the current graph's semantic and degree distributions before
-changing attach thresholds.
-
-Validate the attach quality before relying on a new ANN backend or changed
-threshold:
-
-```bash
-ctx-incremental-shadow \
-  --index-dir ~/.claude/skill-wiki/.embedding-cache/graph/vector-index \
-  --graph ~/.claude/skill-wiki/graphify-out/graph.json \
-  --sample-size 100 \
-  --min-overlap 0.85
-```
-
-The shadow gate pretends sampled existing nodes are new, compares incremental
-attach neighbors to batch graph semantic neighbors, reports precision/recall
-for top 5/10/20, score deltas, and bad examples, then exits non-zero if
-recall at the largest top-k is below the overlap floor.
-
-## Repair Incremental Attach
-
-If the worker says `incremental attach skipped (no vector index)`, build the
-persisted semantic index once:
-
-```bash
-ctx-wiki-graphify \
-  --wiki-dir ~/.claude/skill-wiki \
-  --incremental \
-  --graph-only \
-  --semantic-vector-index numpy-flat
-```
-
-`numpy-flat` is exact and portable. `--semantic-vector-index auto` keeps the
-portable exact backend at this graph size and can switch to the optional ANN
-backend only above the configured node threshold. `hnswlib` is optional and
-should be shadow-gated before release use.
-
-Then process pending entity updates:
-
-```bash
-ctx-wiki-worker --wiki ~/.claude/skill-wiki
-```
-
-That is the supported "attach pending" flow today: the queue is durable, so
-failed or skipped entity-upsert jobs remain visible to the worker and can be
-retried after the index exists. Use manual `ctx-incremental-attach attach
---dry-run` for one-off debugging, not as the normal bulk path.
-
-## Security and Cyber Check
-
-Run this before applying `--update-existing`, before installing a harness with
-approved commands, and before shipping a refreshed graph tarball.
-
-- Inspect changed entity markdown and frontmatter for shell commands, setup
-  commands, install commands, URLs, requested permissions, and model/provider
-  access.
-- Treat these as manual-review blockers: `curl | sh`, `wget | bash`,
-  `Invoke-Expression`, broad `rm -rf`, `git reset --hard`, `chmod 777`, secret
-  upload, disabled auth/TLS/sandboxing/audit/tests, or unpinned package sources.
-- For MCP and harness updates, check network access, filesystem scope, auth
-  material, command transports, and whether setup or verify commands execute
-  remote code.
-- Prefer dry-run first: `ctx-harness-install <slug> --dry-run` and
-  `ctx-harness-install <slug> --update --dry-run`.
-- If a candidate is useful but risky, document the safer install path or keep it
-  as metadata instead of shipping it as an installed skill.
-
-## Updating an Existing Entity
-
-The add commands are non-destructive by default when the target skill, agent,
-MCP server, or harness already exists. The first add attempt prints an update
-review instead of replacing files. That review lists changed fields, expected
-benefits, possible regressions, security findings, and a recommendation.
-
-Use this flow for every entity type:
-
-1. Run the normal add command.
-2. If ctx prints `Existing <type> already exists`, read the benefits and risks.
-3. Keep the current entity by doing nothing, or re-run with `--skip-existing`
-   in batch jobs where you do not want reviews.
-4. Apply the replacement only after review with `--update-existing`.
-5. Drain the queue with `ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1`
-   for immediate local recommendation use, or rebuild with `ctx-wiki-graphify`
-   when the update should be reconciled into shipped graph artifacts.
-
-Examples:
-
-```bash
-ctx-skill-add --skill-path ./SKILL.md --name fastapi-review
-ctx-skill-add --skill-path ./SKILL.md --name fastapi-review --update-existing
-
-ctx-agent-add --agent-path ./code-reviewer.md --name code-reviewer
-ctx-agent-add --agent-path ./code-reviewer.md --name code-reviewer --update-existing
-
-ctx-mcp-add --from-json ./github-mcp.json
-ctx-mcp-add --from-json ./github-mcp.json --update-existing
-
-ctx-harness-add --from-json ./text-to-cad-harness.json
-ctx-harness-add --from-json ./text-to-cad-harness.json --update-existing
-```
-
-`ctx-harness-install --update` is different: it refreshes an installed harness
-checkout under `~/.claude/harnesses/<slug>`. Catalog entity replacement uses
-`ctx-harness-add --update-existing`.
-
-## Removing or Retiring an Entity
-
-Removal has three separate meanings. First decide which one you need.
-
-- **Catalog removal** stops ctx from showing the entity in the wiki, graph, and
-  recommendations. Use the dashboard: `ctx-monitor serve`, open **Manage**,
-  search for the slug and type, then choose **Delete selected**. This deletes
-  the wiki page and queues an `entity-upsert` delete plus `graph-export` job.
-- **Runtime unload** removes a currently loaded entity from the live manifest.
-  Use the dashboard **Loaded** page. MCP unloads call the Claude MCP removal
-  path when available; skill and agent unloads remove the manifest row.
-- **Installed-file removal** is type-specific. Use
-  `ctx-harness-install <slug> --uninstall` for harness checkouts,
-  `ctx-mcp-uninstall <slug>` for installed MCPs, and `ctx-lifecycle archive`
-  then `ctx-lifecycle purge` for stale local skills that should be deleted
-  after the configured grace period.
-
-After deleting an entity page, drain or rebuild before trusting recommendation
-results:
-
-```bash
-ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1
-ctx-scan-repo --repo . --recommend
-```
-
-If you need an auditable manual flow without the browser, use the dashboard
-local API only from loopback with the per-process monitor token printed into
-the served page. The API route is `POST /api/entity/delete` with
-`{"slug": "...", "entity_type": "skill|agent|mcp-server|harness"}`.
-
-## Add a Skill
-
-Use this when you have a local `SKILL.md` that should be installed under
-`~/.claude/skills/<name>/SKILL.md` and mirrored into the wiki.
-
-```bash
-ctx-skill-add \
-  --skill-path ./SKILL.md \
-  --name fastapi-review
-```
-
-What happens:
-
-1. The name is validated.
-2. Intake checks run against the markdown.
-3. The skill is copied into `~/.claude/skills/`.
-4. A wiki page is created under `entities/skills/`.
-5. The wiki index and log are updated.
-
-## Add an Agent
-
-Use this when you have a local Claude Code agent markdown file.
-
-```bash
-ctx-agent-add \
-  --agent-path ./code-reviewer.md \
-  --name code-reviewer
-```
-
-Batch-add every top-level `.md` file in a directory:
-
-```bash
-ctx-agent-add --scan-dir ./agents --skip-existing
-```
-
-Agents are copied into `~/.claude/agents/` and mirrored into
-`entities/agents/`. Re-run `ctx-wiki-graphify` after adding agents if you want
-graph recommendations to include them.
-
-## Add an MCP Server
-
-Use this when you want the MCP server available as a recommendation before
-installing it into a host.
-
-Create `github-mcp.json`:
-
-```json
-{
-  "name": "GitHub MCP",
-  "slug": "github-mcp",
-  "description": "MCP server for GitHub repository and issue workflows.",
-  "github_url": "https://github.com/modelcontextprotocol/servers",
-  "sources": ["manual"],
-  "tags": ["github", "automation", "repository"],
-  "transports": ["stdio"]
-}
-```
-
-Add it:
-
-```bash
-ctx-mcp-add --from-json ./github-mcp.json
-```
-
-MCP pages live under `entities/mcp-servers/<shard>/<slug>.md`. The add command
-detects existing pages by slug and, when possible, canonical GitHub URL. If a
-match exists, ctx prints the update review and skips replacement unless
-`--update-existing` is passed.
-
-## Add a Harness
-
-Use this when a repo provides the runtime around a model rather than just a
-tool. Harness examples include coding-agent loops, CAD-generation runtimes,
-browser-automation runners, evaluation loops, and local-model workbenches.
-
-Example: add `earthtojake/text-to-cad` as a harness recommendation.
-
-```bash
-ctx-harness-add \
-  --repo https://github.com/earthtojake/text-to-cad \
-  --name "Text to CAD" \
-  --description "Harness for turning text prompts into CAD artifacts." \
-  --tag cad --tag 3d --tag automation \
-  --model-provider openai \
-  --runtime python \
-  --capability "Generate CAD artifacts from natural language" \
-  --setup-command "pip install -e ." \
-  --verify-command "pytest"
-```
-
-Or load one JSON record:
-
-```json
-{
-  "repo_url": "https://github.com/earthtojake/text-to-cad",
-  "name": "Text to CAD",
-  "description": "Harness for turning text prompts into CAD artifacts.",
-  "tags": ["cad", "3d", "automation"],
-  "model_providers": ["openai"],
-  "runtimes": ["python"],
-  "capabilities": ["Generate CAD artifacts from natural language"],
-  "setup_commands": ["pip install -e ."],
-  "verify_commands": ["pytest"],
-  "sources": ["manual"]
-}
-```
-
-```bash
-ctx-harness-add --from-json ./text-to-cad-harness.json
-```
-
-Harness pages live under `entities/harnesses/<slug>.md`. Setup and verification
-commands are documentation only; ctx records them so the user can inspect and
-decide before running anything.
-
-To inspect and install a harness:
-
-```bash
-ctx-harness-install text-to-cad --dry-run
-ctx-harness-install text-to-cad
-ctx-harness-install text-to-cad --update --dry-run
-ctx-harness-install text-to-cad --uninstall --dry-run
-```
-
-The installer clones or copies the harness into `~/.claude/harnesses/<slug>` and
-writes `~/.claude/harness-installs/<slug>.json`. It does not run setup commands
-unless you pass `--approve-commands`, and it does not run verification commands
-unless you also pass `--run-verify`.
-
-```bash
-ctx-harness-install text-to-cad --approve-commands --run-verify
-ctx-harness-install text-to-cad --update --approve-commands --run-verify
-ctx-harness-install text-to-cad --uninstall
-ctx-harness-install text-to-cad --uninstall --keep-files
-```
-
-## Initialize Model Choice
-
-During setup, record whether you use Claude Code or your own model. Plain
-`ctx-init` starts a small wizard when it is attached to an interactive
-terminal; use `ctx-init --wizard` to force the prompts, or pass explicit flags
-such as `--model-mode skip` for non-interactive automation.
-
-```bash
-ctx-init
-ctx-init --wizard
-ctx-init --model-mode skip
-```
-
-For Claude Code:
-
-```bash
-ctx-init --model-mode claude-code --goal "maintain a FastAPI service"
-```
-
-For a custom model:
-
-```bash
-ctx-init \
-  --model-mode custom \
-  --model openai/gpt-5.5 \
-  --goal "build CAD artifacts from text prompts"
-```
-
-Add `--validate-model` only when you want `ctx-init` to make one small provider
-call. Without that flag, setup writes `~/.claude/ctx-model-profile.json` and
-prints harness recommendations without calling the model.
+# Entity Onboarding
+
+ctx treats skills, agents, MCP servers, and harnesses as wiki entities that can
+be indexed, linked in the knowledge graph, and recommended from the same
+surface. The important distinction is install behavior:
+
+- Skills and agents are local Claude Code assets.
+- MCP servers are recorded first, then installed only when the user opts in.
+- Harnesses are recorded first. A harness describes the machinery around the
+  model: runtime, tools, access boundaries, memory, verification, and approval
+  policy. Adding one never executes upstream setup commands.
+
+After adding any entity, drain the durable wiki queue when you want the local
+runtime graph to see it immediately:
+
+```bash
+ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1
+ctx-scan-repo --repo . --recommend
+```
+
+If a persisted semantic vector index exists, that worker pass also runs a
+best-effort incremental attach into `graphify-out/entity-overlays.jsonl` so the
+new entity can connect to existing graph nodes without a full all-pairs
+semantic rebuild. The wiki page remains the source of truth; if incremental
+attach is skipped or fails, the worker still queues the normal incremental
+graph export job.
+
+## Updating the Graph and LLM Wiki
+
+Use this sequence for every accepted skill, agent, MCP server, or harness
+change. The graph and LLM-wiki are shippable artifacts, not scratch output, so
+the update is treated like a release step.
+
+1. Add or update the entity through the matching command:
+   `ctx-skill-add`, `ctx-agent-add`, `ctx-mcp-add`, or `ctx-harness-add`.
+2. If the entity already exists, read the update review. It lists changed
+   fields, likely benefits, regressions, and security findings. Do not pass
+   `--update-existing` until those findings are acceptable.
+3. Run the security/cyber check below.
+4. Park heavyweight graph artifacts locally before rebuilds:
+   `python scripts/graph_artifact_guard.py park`. This keeps background Git
+   integrations from repeatedly LFS-cleaning the full wiki tarball while it is
+   still changing.
+5. Drain the wiki queue for local runtime use:
+   `ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1`. This updates the
+   wiki index, attempts incremental ANN graph attach when a vector index exists,
+   and queues the normal incremental graph export.
+6. Rebuild the curated wiki graph with `ctx-wiki-graphify` before shipping
+   release artifacts or when you need a full graph/export reconciliation.
+7. Repack `graph/wiki-graph.tar.gz` through the artifact promotion path:
+   write a staged tarball, validate it, atomically promote it, and keep the
+   generated `*.promotion.json` metadata with the previous/current hashes.
+   Never commit local review reports or raw caches.
+8. Refresh the bulk skill index when shipping large skill updates.
+   This adds first-class `skill` nodes, skill pages under `entities/skills/`, install
+   commands, duplicate hints, and metadata-only quality/security signals:
+
+   ```bash
+   python src/import_skills_sh_catalog.py --from-api-union <raw.json> \
+     --catalog-out graph/<skill-index>.json.gz \
+     --wiki-tar graph/wiki-graph.tar.gz \
+     --update-wiki-tar
+   ```
+9. Refresh published counts with `python src/update_repo_stats.py`.
+10. Verify the changed entity can be recommended through
+   `ctx-scan-repo --repo . --recommend` or `ctx__recommend_bundle`.
+11. Unpark and stage the graph artifacts once the release candidate is final:
+    `python scripts/graph_artifact_guard.py unpark`, then `git add` the graph
+    artifacts intentionally. Run `python scripts/graph_artifact_guard.py prune`
+    after interrupted Git/LFS runs or after release staging to clean prunable
+    local LFS cache entries. Add `--include-git-prune` only when you explicitly
+    want repo-wide dangling Git objects removed too.
+
+The durable wiki worker drains `entity-upsert`, `graph-export`,
+`skill-index-refresh`, `tar-refresh`, and `artifact-promotion` jobs. Use
+`ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1` for a controlled
+single-job drain, or omit `--limit` to drain the ready queue.
+
+For a manual attach dry-run against an existing vector index:
+
+```bash
+ctx-incremental-attach attach \
+  --index-dir ~/.claude/skill-wiki/.embedding-cache/graph/vector-index \
+  --overlay ~/.claude/skill-wiki/graphify-out/entity-overlays.jsonl \
+  --node-id skill:fastapi-review \
+  --type skill \
+  --label fastapi-review \
+  --text-file ~/.claude/skill-wiki/entities/skills/fastapi-review.md \
+  --dry-run
+```
+
+Use `ctx-incremental-attach calibrate --graph ~/.claude/skill-wiki/graphify-out/graph.json`
+to inspect the current graph's semantic and degree distributions before
+changing attach thresholds.
+
+Validate the attach quality before relying on a new ANN backend or changed
+threshold:
+
+```bash
+ctx-incremental-shadow \
+  --index-dir ~/.claude/skill-wiki/.embedding-cache/graph/vector-index \
+  --graph ~/.claude/skill-wiki/graphify-out/graph.json \
+  --sample-size 100 \
+  --min-overlap 0.85
+```
+
+The shadow gate pretends sampled existing nodes are new, compares incremental
+attach neighbors to batch graph semantic neighbors, reports precision/recall
+for top 5/10/20, score deltas, and bad examples, then exits non-zero if
+recall at the largest top-k is below the overlap floor.
+
+## Repair Incremental Attach
+
+If the worker says `incremental attach skipped (no vector index)`, build the
+persisted semantic index once:
+
+```bash
+ctx-wiki-graphify \
+  --wiki-dir ~/.claude/skill-wiki \
+  --incremental \
+  --graph-only \
+  --semantic-vector-index numpy-flat
+```
+
+`numpy-flat` is exact and portable. `--semantic-vector-index auto` keeps the
+portable exact backend at this graph size and can switch to the optional ANN
+backend only above the configured node threshold. `hnswlib` is optional and
+should be shadow-gated before release use.
+
+Then process pending entity updates:
+
+```bash
+ctx-wiki-worker --wiki ~/.claude/skill-wiki
+```
+
+That is the supported "attach pending" flow today: the queue is durable, so
+failed or skipped entity-upsert jobs remain visible to the worker and can be
+retried after the index exists. Use manual `ctx-incremental-attach attach
+--dry-run` for one-off debugging, not as the normal bulk path.
+
+## Security and Cyber Check
+
+Run this before applying `--update-existing`, before installing a harness with
+approved commands, and before shipping a refreshed graph tarball.
+
+- Inspect changed entity markdown and frontmatter for shell commands, setup
+  commands, install commands, URLs, requested permissions, and model/provider
+  access.
+- Treat these as manual-review blockers: `curl | sh`, `wget | bash`,
+  `Invoke-Expression`, broad `rm -rf`, `git reset --hard`, `chmod 777`, secret
+  upload, disabled auth/TLS/sandboxing/audit/tests, or unpinned package sources.
+- For MCP and harness updates, check network access, filesystem scope, auth
+  material, command transports, and whether setup or verify commands execute
+  remote code.
+- Prefer dry-run first: `ctx-harness-install <slug> --dry-run` and
+  `ctx-harness-install <slug> --update --dry-run`.
+- If a candidate is useful but risky, document the safer install path or keep it
+  as metadata instead of shipping it as an installed skill.
+
+## Updating an Existing Entity
+
+The add commands are non-destructive by default when the target skill, agent,
+MCP server, or harness already exists. The first add attempt prints an update
+review instead of replacing files. That review lists changed fields, expected
+benefits, possible regressions, security findings, and a recommendation.
+
+Use this flow for every entity type:
+
+1. Run the normal add command.
+2. If ctx prints `Existing <type> already exists`, read the benefits and risks.
+3. Keep the current entity by doing nothing, or re-run with `--skip-existing`
+   in batch jobs where you do not want reviews.
+4. Apply the replacement only after review with `--update-existing`.
+5. Drain the queue with `ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1`
+   for immediate local recommendation use, or rebuild with `ctx-wiki-graphify`
+   when the update should be reconciled into shipped graph artifacts.
+
+Examples:
+
+```bash
+ctx-skill-add --skill-path ./SKILL.md --name fastapi-review
+ctx-skill-add --skill-path ./SKILL.md --name fastapi-review --update-existing
+
+ctx-agent-add --agent-path ./code-reviewer.md --name code-reviewer
+ctx-agent-add --agent-path ./code-reviewer.md --name code-reviewer --update-existing
+
+ctx-mcp-add --from-json ./github-mcp.json
+ctx-mcp-add --from-json ./github-mcp.json --update-existing
+
+ctx-harness-add --from-json ./text-to-cad-harness.json
+ctx-harness-add --from-json ./text-to-cad-harness.json --update-existing
+```
+
+`ctx-harness-install --update` is different: it refreshes an installed harness
+checkout under `~/.claude/harnesses/<slug>`. Catalog entity replacement uses
+`ctx-harness-add --update-existing`.
+
+## Removing or Retiring an Entity
+
+Removal has three separate meanings. First decide which one you need.
+
+- **Catalog removal** stops ctx from showing the entity in the wiki, graph, and
+  recommendations. Use the dashboard: `ctx-monitor serve`, open **Manage**,
+  search for the slug and type, then choose **Delete selected**. This deletes
+  the wiki page and queues an `entity-upsert` delete plus `graph-export` job.
+- **Runtime unload** removes a currently loaded entity from the live manifest.
+  Use the dashboard **Loaded** page. MCP unloads call the Claude MCP removal
+  path when available; skill and agent unloads remove the manifest row.
+- **Installed-file removal** is type-specific. Use
+  `ctx-harness-install <slug> --uninstall` for harness checkouts,
+  `ctx-mcp-uninstall <slug>` for installed MCPs, and `ctx-lifecycle archive`
+  then `ctx-lifecycle purge` for stale local skills that should be deleted
+  after the configured grace period.
+
+After deleting an entity page, drain or rebuild before trusting recommendation
+results:
+
+```bash
+ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1
+ctx-scan-repo --repo . --recommend
+```
+
+If you need an auditable manual flow without the browser, use the dashboard
+local API only from loopback with the per-process monitor token printed into
+the served page. The API route is `POST /api/entity/delete` with
+`{"slug": "...", "entity_type": "skill|agent|mcp-server|harness"}`.
+
+## Add a Skill
+
+Use this when you have a local `SKILL.md` that should be installed under
+`~/.claude/skills/<name>/SKILL.md` and mirrored into the wiki.
+
+```bash
+ctx-skill-add \
+  --skill-path ./SKILL.md \
+  --name fastapi-review
+```
+
+What happens:
+
+1. The name is validated.
+2. Intake checks run against the markdown.
+3. The skill is copied into `~/.claude/skills/`.
+4. A wiki page is created under `entities/skills/`.
+5. The wiki index and log are updated.
+
+## Add an Agent
+
+Use this when you have a local Claude Code agent markdown file.
+
+```bash
+ctx-agent-add \
+  --agent-path ./code-reviewer.md \
+  --name code-reviewer
+```
+
+Batch-add every top-level `.md` file in a directory:
+
+```bash
+ctx-agent-add --scan-dir ./agents --skip-existing
+```
+
+Agents are copied into `~/.claude/agents/` and mirrored into
+`entities/agents/`. Re-run `ctx-wiki-graphify` after adding agents if you want
+graph recommendations to include them.
+
+## Add an MCP Server
+
+Use this when you want the MCP server available as a recommendation before
+installing it into a host.
+
+Create `github-mcp.json`:
+
+```json
+{
+  "name": "GitHub MCP",
+  "slug": "github-mcp",
+  "description": "MCP server for GitHub repository and issue workflows.",
+  "github_url": "https://github.com/modelcontextprotocol/servers",
+  "sources": ["manual"],
+  "tags": ["github", "automation", "repository"],
+  "transports": ["stdio"]
+}
+```
+
+Add it:
+
+```bash
+ctx-mcp-add --from-json ./github-mcp.json
+```
+
+MCP pages live under `entities/mcp-servers/<shard>/<slug>.md`. The add command
+detects existing pages by slug and, when possible, canonical GitHub URL. If a
+match exists, ctx prints the update review and skips replacement unless
+`--update-existing` is passed.
+
+## Add a Harness
+
+Use this when a repo provides the runtime around a model rather than just a
+tool. Harness examples include coding-agent loops, CAD-generation runtimes,
+browser-automation runners, evaluation loops, and local-model workbenches.
+
+Example: add `earthtojake/text-to-cad` as a harness recommendation.
+
+```bash
+ctx-harness-add \
+  --repo https://github.com/earthtojake/text-to-cad \
+  --name "Text to CAD" \
+  --description "Harness for turning text prompts into CAD artifacts." \
+  --tag cad --tag 3d --tag automation \
+  --model-provider openai \
+  --runtime python \
+  --capability "Generate CAD artifacts from natural language" \
+  --setup-command "pip install -e ." \
+  --verify-command "pytest"
+```
+
+Or load one JSON record:
+
+```json
+{
+  "repo_url": "https://github.com/earthtojake/text-to-cad",
+  "name": "Text to CAD",
+  "description": "Harness for turning text prompts into CAD artifacts.",
+  "tags": ["cad", "3d", "automation"],
+  "model_providers": ["openai"],
+  "runtimes": ["python"],
+  "capabilities": ["Generate CAD artifacts from natural language"],
+  "setup_commands": ["pip install -e ."],
+  "verify_commands": ["pytest"],
+  "sources": ["manual"]
+}
+```
+
+```bash
+ctx-harness-add --from-json ./text-to-cad-harness.json
+```
+
+Harness pages live under `entities/harnesses/<slug>.md`. Setup and verification
+commands are documentation only; ctx records them so the user can inspect and
+decide before running anything.
+
+To inspect and install a harness:
+
+```bash
+ctx-harness-install text-to-cad --dry-run
+ctx-harness-install text-to-cad
+ctx-harness-install text-to-cad --update --dry-run
+ctx-harness-install text-to-cad --uninstall --dry-run
+```
+
+The installer clones or copies the harness into `~/.claude/harnesses/<slug>` and
+writes `~/.claude/harness-installs/<slug>.json`. It does not run setup commands
+unless you pass `--approve-commands`, and it does not run verification commands
+unless you also pass `--run-verify`.
+
+```bash
+ctx-harness-install text-to-cad --approve-commands --run-verify
+ctx-harness-install text-to-cad --update --approve-commands --run-verify
+ctx-harness-install text-to-cad --uninstall
+ctx-harness-install text-to-cad --uninstall --keep-files
+```
+
+## Initialize Model Choice
+
+During setup, record whether you use Claude Code or your own model. Plain
+`ctx-init` starts a small wizard when it is attached to an interactive
+terminal; use `ctx-init --wizard` to force the prompts, or pass explicit flags
+such as `--model-mode skip` for non-interactive automation.
+
+```bash
+ctx-init
+ctx-init --wizard
+ctx-init --model-mode skip
+```
+
+For Claude Code:
+
+```bash
+ctx-init --model-mode claude-code --goal "maintain a FastAPI service"
+```
+
+For a custom model:
+
+```bash
+ctx-init \
+  --model-mode custom \
+  --model openai/gpt-5.5 \
+  --goal "build CAD artifacts from text prompts"
+```
+
+Add `--validate-model` only when you want `ctx-init` to make one small provider
+call. Without that flag, setup writes `~/.claude/ctx-model-profile.json` and
+prints harness recommendations without calling the model.

docs/harness/attaching-to-hosts.md

@@ -1,289 +1,289 @@
-# Attaching ctx to any LLM host
-
-`ctx` ships three integration surfaces. Pick based on what your host
-already supports:
-
-| Your host | Use |
-|---|---|
-| MCP-native (Claude Code, Claude Agent SDK, Cline, Goose, OpenHands, Continue) | **MCP server** — no Python, just spawn `ctx-mcp-server` |
-| Anything that isn't MCP-native but runs Python | **Python library** — `from ctx import recommend_bundle, ...` |
-| "I just want to run an agent and get recommendations" | **`ctx run` CLI** — our built-in harness |
-
-All three paths consume the **same** knowledge graph, llm-wiki, and
-quality scoring. Recommendations are identical; only the transport
-differs.
-
----
-
-## 1. MCP server path
-
-Install ctx with the harness extras:
-
-```bash
-pip install "claude-ctx[harness]"
-```
-
-This puts `ctx-mcp-server` on your PATH. Then wire it into your host:
-
-### Claude Code
-
-```bash
-claude mcp add ctx-wiki -- ctx-mcp-server
-```
-
-The tools `ctx__recommend_bundle`, `ctx__graph_query`, `ctx__wiki_search`,
-`ctx__wiki_get` appear to Claude on the next turn. Ask
-"What skills help with FastAPI auth?" and it will call them.
-
-### Claude Agent SDK (Python)
-
-```python
-from anthropic import Anthropic
-from claude_agent_sdk import ClaudeAgentOptions, McpServerConfig
-
-options = ClaudeAgentOptions(
-    mcp_servers={
-        "ctx-wiki": McpServerConfig(
-            command="ctx-mcp-server",
-        ),
-    },
-)
-```
-
-### Cline / Continue.dev
-
-Add to your MCP server config (`~/.config/cline/mcp.json` or the
-Continue equivalent):
-
-```json
-{
-  "mcpServers": {
-    "ctx-wiki": {
-      "command": "ctx-mcp-server"
-    }
-  }
-}
-```
-
-### Goose
-
-`~/.config/goose/config.yaml`:
-
-```yaml
-extensions:
-  ctx-wiki:
-    type: stdio
-    cmd: ctx-mcp-server
-```
-
-### OpenHands
-
-OpenHands' runtime config:
-
-```json
-{
-  "mcp_servers": {
-    "ctx-wiki": {
-      "command": "ctx-mcp-server"
-    }
-  }
-}
-```
-
-### Any MCP-speaking harness
-
-The server reads JSON-RPC 2.0 on stdin, writes on stdout, speaks
-MCP protocol version `2024-11-05`. Any client that does the standard
-`initialize` handshake + `tools/list` + `tools/call` flow works.
-
-### Live MCP compatibility gate
-
-The regular test suite never starts arbitrary third-party MCP servers.
-Those commands run as local subprocesses and can read files, use the
-network, and inherit whatever environment you explicitly allow.
-
-To validate a trusted server, provide a local config and opt in:
-
-```bash
-python -m pytest src/tests/test_mcp_live_compat.py \
-  --run-live-mcp \
-  --live-mcp-config /path/to/trusted-mcp.json
-```
-
-Example config:
-
-```json
-{
-  "name": "trusted-filesystem",
-  "command": "npx",
-  "args": ["-y", "@modelcontextprotocol/server-filesystem", "${tmp_path}"],
-  "startup_timeout": 30,
-  "request_timeout": 10,
-  "inherit_env": false,
-  "env": {},
-  "expected_tools": ["list_directory"],
-  "probe": {
-    "tool": "list_directory",
-    "arguments": {"path": "."},
-    "expect_text_contains": ""
-  },
-  "trust": {
-    "server_is_third_party_code": true,
-    "approved_by": "your-name"
-  }
-}
-```
-
-`command` and `args` are passed as an argv list, not through a shell.
-Parent secrets are not inherited unless you set `inherit_env: true`; prefer
-explicit `env` keys for servers that need credentials. `${tmp_path}` expands
-to a pytest temporary directory so filesystem probes can avoid real user data.
-
----
-
-## 2. Python library path
-
-For custom harnesses that aren't MCP-native but can import Python:
-
-```python
-from ctx import (
-    recommend_bundle,   # free-text → ranked skill/agent/MCP bundle
-    graph_query,        # walk from seed entities
-    wiki_search,        # keyword search entity pages
-    wiki_get,           # fetch one entity by slug
-    list_all_entities,  # enumerate every slug
-)
-
-# Inside your agent loop:
-def on_user_turn(query: str):
-    bundle = recommend_bundle(query, top_k=5)
-    for entry in bundle:
-        print(f"  [{entry['type']:>11}] {entry['name']}  (score {entry['score']:.1f})")
-
-    # User asks about a specific slug you saw in the bundle:
-    page = wiki_get("fastapi-pro")
-    if page:
-        inject_into_context(page["body"])
-```
-
-The first call to any of these lazy-loads the graph + wiki once;
-subsequent calls are O(walk) cheap. Safe to call from inside your
-own while-loop on every turn.
-
-Advanced: build a `CtxCoreToolbox` directly if you need to point at
-a non-default wiki/graph path:
-
-```python
-from pathlib import Path
-from ctx import CtxCoreToolbox
-
-toolbox = CtxCoreToolbox(
-    wiki_dir=Path("/path/to/custom/wiki"),
-    graph_path=Path("/path/to/custom/graph.json"),
-)
-for td in toolbox.tool_definitions():
-    print(td.name, td.description[:50])
-```
-
----
-
-## 3. `ctx run` CLI path
-
-If you don't have your own loop yet:
-
-```bash
-pip install "claude-ctx[harness]"
-export OPENROUTER_API_KEY=sk-or-v1-...
-
-ctx run \
-    --model openrouter/anthropic/claude-opus-4.7 \
-    --task "find the failing tests in this repo and fix them" \
-    --mcp filesystem \
-    --budget-usd 2.00
-```
-
-Or offline with Ollama:
-
-```bash
-ctx run \
-    --model ollama/llama3.1:70b \
-    --task "summarize the architecture" \
-    --mcp filesystem
-```
-
-See `ctx run --help` for the full flag set (budgets, compaction,
-system prompt overrides, session resume, JSON output, ...).
-
----
-
-## Installed harness attachment
-
-`ctx-harness-install <slug>` creates `.ctx/attach/` inside the installed
-harness target. The directory contains the attach files for the modes that
-catalog entry supports:
-
-- `README.md` describes the supported modes and safety expectations.
-- `mcp.json` starts `ctx-mcp-server` for MCP-speaking hosts.
-- `python.py` shows the Python recommendation/wiki calls for custom loops.
-- `ctx-run.txt` gives a `ctx run` command template.
-
-The install command does not run the harness or store secrets in those files.
-Setup commands still require `--approve-commands`; verification commands still
-require `--run-verify`.
-
-If no catalog harness fits, generate a build handoff instead of forcing a weak
-match:
-
-```bash
-ctx-harness-install --recommend \
-  --goal "build a private CAD workflow with a local model" \
-  --model-provider ollama \
-  --model ollama/llama3.1 \
-  --plan-on-no-fit \
-  --plan-output custom-harness.md
-```
-
----
-
-## Choosing the right path
-
-| Situation | Path |
-|---|---|
-| Your host already speaks MCP | 1 (MCP server) — zero Python code on your side |
-| You want the alive-skill system inside your existing Python loop | 2 (library) |
-| You're comparing models and need a harness | 3 (CLI) |
-| No catalog harness fits your model/goal | generated custom harness plan |
-| You're building an IDE extension | 1 if the IDE speaks MCP (most do), else 2 |
-
-All three paths share `~/.claude/skill-wiki/` as the source-of-truth
-corpus, so your recommendations are consistent regardless of the
-integration you pick.
-
----
-
-## Skill lifecycle
-
-Recommendations go up and down based on use automatically. `ctx`
-tracks:
-
-- **How recently a skill was invoked** (`telemetry_signal`).
-- **How broadly it's used across the graph** (`graph_signal`).
-- **Whether new skills are being added** (`intake_signal`).
-
-Skills that fall below a quality floor get demoted to `stale` status
-and de-ranked from future recommendations. This logic lives in
-`ctx.core.quality.quality_signals` and runs identically whether
-you're on the MCP path, library path, or `ctx run` CLI.
-
-To inspect lifecycle state for a specific skill:
-
-```bash
-ctx-skill-quality explain fastapi-pro
-```
-
-Or from Python:
-
-```python
-from ctx.core.quality import quality_signals
-# see ctx.core.quality for the scoring API
-```
+# Attaching ctx to any LLM host
+
+`ctx` ships three integration surfaces. Pick based on what your host
+already supports:
+
+| Your host | Use |
+|---|---|
+| MCP-native (Claude Code, Claude Agent SDK, Cline, Goose, OpenHands, Continue) | **MCP server** — no Python, just spawn `ctx-mcp-server` |
+| Anything that isn't MCP-native but runs Python | **Python library** — `from ctx import recommend_bundle, ...` |
+| "I just want to run an agent and get recommendations" | **`ctx run` CLI** — our built-in harness |
+
+All three paths consume the **same** knowledge graph, llm-wiki, and
+quality scoring. Recommendations are identical; only the transport
+differs.
+
+---
+
+## 1. MCP server path
+
+Install ctx with the harness extras:
+
+```bash
+pip install "claude-ctx[harness]"
+```
+
+This puts `ctx-mcp-server` on your PATH. Then wire it into your host:
+
+### Claude Code
+
+```bash
+claude mcp add ctx-wiki -- ctx-mcp-server
+```
+
+The tools `ctx__recommend_bundle`, `ctx__graph_query`, `ctx__wiki_search`,
+`ctx__wiki_get` appear to Claude on the next turn. Ask
+"What skills help with FastAPI auth?" and it will call them.
+
+### Claude Agent SDK (Python)
+
+```python
+from anthropic import Anthropic
+from claude_agent_sdk import ClaudeAgentOptions, McpServerConfig
+
+options = ClaudeAgentOptions(
+    mcp_servers={
+        "ctx-wiki": McpServerConfig(
+            command="ctx-mcp-server",
+        ),
+    },
+)
+```
+
+### Cline / Continue.dev
+
+Add to your MCP server config (`~/.config/cline/mcp.json` or the
+Continue equivalent):
+
+```json
+{
+  "mcpServers": {
+    "ctx-wiki": {
+      "command": "ctx-mcp-server"
+    }
+  }
+}
+```
+
+### Goose
+
+`~/.config/goose/config.yaml`:
+
+```yaml
+extensions:
+  ctx-wiki:
+    type: stdio
+    cmd: ctx-mcp-server
+```
+
+### OpenHands
+
+OpenHands' runtime config:
+
+```json
+{
+  "mcp_servers": {
+    "ctx-wiki": {
+      "command": "ctx-mcp-server"
+    }
+  }
+}
+```
+
+### Any MCP-speaking harness
+
+The server reads JSON-RPC 2.0 on stdin, writes on stdout, speaks
+MCP protocol version `2024-11-05`. Any client that does the standard
+`initialize` handshake + `tools/list` + `tools/call` flow works.
+
+### Live MCP compatibility gate
+
+The regular test suite never starts arbitrary third-party MCP servers.
+Those commands run as local subprocesses and can read files, use the
+network, and inherit whatever environment you explicitly allow.
+
+To validate a trusted server, provide a local config and opt in:
+
+```bash
+python -m pytest src/tests/test_mcp_live_compat.py \
+  --run-live-mcp \
+  --live-mcp-config /path/to/trusted-mcp.json
+```
+
+Example config:
+
+```json
+{
+  "name": "trusted-filesystem",
+  "command": "npx",
+  "args": ["-y", "@modelcontextprotocol/server-filesystem", "${tmp_path}"],
+  "startup_timeout": 30,
+  "request_timeout": 10,
+  "inherit_env": false,
+  "env": {},
+  "expected_tools": ["list_directory"],
+  "probe": {
+    "tool": "list_directory",
+    "arguments": {"path": "."},
+    "expect_text_contains": ""
+  },
+  "trust": {
+    "server_is_third_party_code": true,
+    "approved_by": "your-name"
+  }
+}
+```
+
+`command` and `args` are passed as an argv list, not through a shell.
+Parent secrets are not inherited unless you set `inherit_env: true`; prefer
+explicit `env` keys for servers that need credentials. `${tmp_path}` expands
+to a pytest temporary directory so filesystem probes can avoid real user data.
+
+---
+
+## 2. Python library path
+
+For custom harnesses that aren't MCP-native but can import Python:
+
+```python
+from ctx import (
+    recommend_bundle,   # free-text → ranked skill/agent/MCP bundle
+    graph_query,        # walk from seed entities
+    wiki_search,        # keyword search entity pages
+    wiki_get,           # fetch one entity by slug
+    list_all_entities,  # enumerate every slug
+)
+
+# Inside your agent loop:
+def on_user_turn(query: str):
+    bundle = recommend_bundle(query, top_k=5)
+    for entry in bundle:
+        print(f"  [{entry['type']:>11}] {entry['name']}  (score {entry['score']:.1f})")
+
+    # User asks about a specific slug you saw in the bundle:
+    page = wiki_get("fastapi-pro")
+    if page:
+        inject_into_context(page["body"])
+```
+
+The first call to any of these lazy-loads the graph + wiki once;
+subsequent calls are O(walk) cheap. Safe to call from inside your
+own while-loop on every turn.
+
+Advanced: build a `CtxCoreToolbox` directly if you need to point at
+a non-default wiki/graph path:
+
+```python
+from pathlib import Path
+from ctx import CtxCoreToolbox
+
+toolbox = CtxCoreToolbox(
+    wiki_dir=Path("/path/to/custom/wiki"),
+    graph_path=Path("/path/to/custom/graph.json"),
+)
+for td in toolbox.tool_definitions():
+    print(td.name, td.description[:50])
+```
+
+---
+
+## 3. `ctx run` CLI path
+
+If you don't have your own loop yet:
+
+```bash
+pip install "claude-ctx[harness]"
+export OPENROUTER_API_KEY=sk-or-v1-...
+
+ctx run \
+    --model openrouter/anthropic/claude-opus-4.7 \
+    --task "find the failing tests in this repo and fix them" \
+    --mcp filesystem \
+    --budget-usd 2.00
+```
+
+Or offline with Ollama:
+
+```bash
+ctx run \
+    --model ollama/llama3.1:70b \
+    --task "summarize the architecture" \
+    --mcp filesystem
+```
+
+See `ctx run --help` for the full flag set (budgets, compaction,
+system prompt overrides, session resume, JSON output, ...).
+
+---
+
+## Installed harness attachment
+
+`ctx-harness-install <slug>` creates `.ctx/attach/` inside the installed
+harness target. The directory contains the attach files for the modes that
+catalog entry supports:
+
+- `README.md` describes the supported modes and safety expectations.
+- `mcp.json` starts `ctx-mcp-server` for MCP-speaking hosts.
+- `python.py` shows the Python recommendation/wiki calls for custom loops.
+- `ctx-run.txt` gives a `ctx run` command template.
+
+The install command does not run the harness or store secrets in those files.
+Setup commands still require `--approve-commands`; verification commands still
+require `--run-verify`.
+
+If no catalog harness fits, generate a build handoff instead of forcing a weak
+match:
+
+```bash
+ctx-harness-install --recommend \
+  --goal "build a private CAD workflow with a local model" \
+  --model-provider ollama \
+  --model ollama/llama3.1 \
+  --plan-on-no-fit \
+  --plan-output custom-harness.md
+```
+
+---
+
+## Choosing the right path
+
+| Situation | Path |
+|---|---|
+| Your host already speaks MCP | 1 (MCP server) — zero Python code on your side |
+| You want the alive-skill system inside your existing Python loop | 2 (library) |
+| You're comparing models and need a harness | 3 (CLI) |
+| No catalog harness fits your model/goal | generated custom harness plan |
+| You're building an IDE extension | 1 if the IDE speaks MCP (most do), else 2 |
+
+All three paths share `~/.claude/skill-wiki/` as the source-of-truth
+corpus, so your recommendations are consistent regardless of the
+integration you pick.
+
+---
+
+## Skill lifecycle
+
+Recommendations go up and down based on use automatically. `ctx`
+tracks:
+
+- **How recently a skill was invoked** (`telemetry_signal`).
+- **How broadly it's used across the graph** (`graph_signal`).
+- **Whether new skills are being added** (`intake_signal`).
+
+Skills that fall below a quality floor get demoted to `stale` status
+and de-ranked from future recommendations. This logic lives in
+`ctx.core.quality.quality_signals` and runs identically whether
+you're on the MCP path, library path, or `ctx run` CLI.
+
+To inspect lifecycle state for a specific skill:
+
+```bash
+ctx-skill-quality explain fastapi-pro
+```
+
+Or from Python:
+
+```python
+from ctx.core.quality import quality_signals
+# see ctx.core.quality for the scoring API
+```

docs/harness/clean-host-contract.md

@@ -1,103 +1,103 @@
-# Clean Host Contract
-
-The clean-host contract is a release-hardening check for ctx. It builds the
-current source tree into a wheel, installs that wheel into a fresh virtualenv,
-redirects user-state environment variables into a temporary directory, and then
-drives real console scripts.
-
-It is intentionally implemented as `scripts/clean_host_contract.py`, not as a
-public `ctx-*` command. The runner is infrastructure for maintainers until the
-contract stabilizes.
-
-## What It Proves
-
-- The source tree can build a wheel.
-- The built wheel installs into a clean virtualenv.
-- Console-script entrypoints execute from the installed wheel.
-- `ctx-init --hooks` writes Claude settings only under an isolated temp home.
-- A deterministic fake Claude host reads the generated settings and executes
-  the installed PostToolUse and Stop hook commands without calling Anthropic
-  APIs.
-- With `--run-live-claude`, a real Claude Code host can be exercised behind an
-  explicit quota acknowledgement, non-spending preflights, and a hard budget
-  cap. Hook execution is verified through a JSONL sentinel written by injected
-  PostToolUse and Stop hooks under the temp root.
-- `ctx-scan-repo --repo <path> --recommend` can scan a tiny FastAPI-like repo from the wheel.
-- `ctx run` can start a session with a process-local fake LiteLLM provider.
-- `ctx resume` can continue that session from the same isolated session store.
-- `--deny-tool` blocks a model-requested ctx tool call before dispatch.
-- Caller `PYTHONPATH` is stripped so the contract cannot accidentally import
-  source-tree modules instead of the installed wheel.
-
-## What It Skips
-
-- It does not run `ctx-init --graph` by default; even the fast runtime graph
-  install downloads or extracts a large release artifact, so it stays in
-  dedicated public-install smoke tests.
-- It does not execute hooks inside a live Claude Code process by default. The
-  live host path is opt-in because it can consume Anthropic or provider quota.
-- It does not connect to a real third-party MCP server.
-- It does not browser-test the monitor dashboard.
-- It does not simulate process kills or power loss during writes.
-
-Those checks stay intentionally manual or opt-in until they are stable enough
-for the default CI path.
-
-## Local Usage
-
-Run from the repository root:
-
-```bash
-python scripts/clean_host_contract.py --fast
-```
-
-For debugging, keep the temp directory:
-
-```bash
-python scripts/clean_host_contract.py --fast --keep-temp
-```
-
-To force a specific temp root:
-
-```bash
-python scripts/clean_host_contract.py --fast --temp-root /tmp/ctx-clean-host-debug
-```
-
-To run the real Claude Code host gate, use a shell with explicit non-file auth
-available, acknowledge quota, and keep the budget small:
-
-```bash
-CTX_LIVE_CLAUDE_ACK=uses_quota \
-python scripts/clean_host_contract.py --fast --run-live-claude --live-claude-max-budget-usd 0.05
-```
-
-Use `--claude-bin /path/to/claude` if `claude` is not on `PATH`. The live gate
-runs `claude --version` and `claude auth status` before the budgeted prompt,
-then appends sentinel hooks to the isolated `settings.json` and requires both
-PostToolUse and Stop records in `live-claude-hooks.jsonl`. It intentionally
-does not read OAuth or keychain state from the real user home; use explicit
-environment/provider auth for this check.
-
-## CI Usage
-
-The main `.github/workflows/test.yml` workflow runs this contract on pushes and
-pull requests. The standalone `.github/workflows/clean-host-contract.yml`
-workflow remains available for manual runs and weekly scheduled drift checks.
-CI uses the default fake-host path and does not spend model quota.
-
-## Failure Triage
-
-- Wheel build failure: inspect package metadata and `pyproject.toml`.
-- Install failure: inspect dependency constraints and `pip check` output.
-- `ctx-init` failure: inspect packaged entrypoints and hook module paths.
-- Fake Claude hook-smoke failure: inspect generated `settings.json`, packaged
-  hook module paths, and whether PostToolUse/Stop hook schemas changed.
-- Live Claude gate failure: inspect explicit auth env/provider credentials,
-  `claude auth status`, the budget cap, the injected sentinel hook entries, and
-  `live-claude-hooks.jsonl` under the temp root.
-- `ctx-scan-repo` failure: inspect installed flat-module entrypoints and
-  resolver imports.
-- `ctx run` or `ctx resume` failure: inspect LiteLLM provider import behavior,
-  session store paths, and CLI metadata replay.
-- Tool denial failure: inspect `--allow-tool`/`--deny-tool` policy handling in
-  `src/ctx/cli/run.py` and `src/ctx/adapters/generic/loop.py`.
+# Clean Host Contract
+
+The clean-host contract is a release-hardening check for ctx. It builds the
+current source tree into a wheel, installs that wheel into a fresh virtualenv,
+redirects user-state environment variables into a temporary directory, and then
+drives real console scripts.
+
+It is intentionally implemented as `scripts/clean_host_contract.py`, not as a
+public `ctx-*` command. The runner is infrastructure for maintainers until the
+contract stabilizes.
+
+## What It Proves
+
+- The source tree can build a wheel.
+- The built wheel installs into a clean virtualenv.
+- Console-script entrypoints execute from the installed wheel.
+- `ctx-init --hooks` writes Claude settings only under an isolated temp home.
+- A deterministic fake Claude host reads the generated settings and executes
+  the installed PostToolUse and Stop hook commands without calling Anthropic
+  APIs.
+- With `--run-live-claude`, a real Claude Code host can be exercised behind an
+  explicit quota acknowledgement, non-spending preflights, and a hard budget
+  cap. Hook execution is verified through a JSONL sentinel written by injected
+  PostToolUse and Stop hooks under the temp root.
+- `ctx-scan-repo --repo <path> --recommend` can scan a tiny FastAPI-like repo from the wheel.
+- `ctx run` can start a session with a process-local fake LiteLLM provider.
+- `ctx resume` can continue that session from the same isolated session store.
+- `--deny-tool` blocks a model-requested ctx tool call before dispatch.
+- Caller `PYTHONPATH` is stripped so the contract cannot accidentally import
+  source-tree modules instead of the installed wheel.
+
+## What It Skips
+
+- It does not run `ctx-init --graph` by default; even the fast runtime graph
+  install downloads or extracts a large release artifact, so it stays in
+  dedicated public-install smoke tests.
+- It does not execute hooks inside a live Claude Code process by default. The
+  live host path is opt-in because it can consume Anthropic or provider quota.
+- It does not connect to a real third-party MCP server.
+- It does not browser-test the monitor dashboard.
+- It does not simulate process kills or power loss during writes.
+
+Those checks stay intentionally manual or opt-in until they are stable enough
+for the default CI path.
+
+## Local Usage
+
+Run from the repository root:
+
+```bash
+python scripts/clean_host_contract.py --fast
+```
+
+For debugging, keep the temp directory:
+
+```bash
+python scripts/clean_host_contract.py --fast --keep-temp
+```
+
+To force a specific temp root:
+
+```bash
+python scripts/clean_host_contract.py --fast --temp-root /tmp/ctx-clean-host-debug
+```
+
+To run the real Claude Code host gate, use a shell with explicit non-file auth
+available, acknowledge quota, and keep the budget small:
+
+```bash
+CTX_LIVE_CLAUDE_ACK=uses_quota \
+python scripts/clean_host_contract.py --fast --run-live-claude --live-claude-max-budget-usd 0.05
+```
+
+Use `--claude-bin /path/to/claude` if `claude` is not on `PATH`. The live gate
+runs `claude --version` and `claude auth status` before the budgeted prompt,
+then appends sentinel hooks to the isolated `settings.json` and requires both
+PostToolUse and Stop records in `live-claude-hooks.jsonl`. It intentionally
+does not read OAuth or keychain state from the real user home; use explicit
+environment/provider auth for this check.
+
+## CI Usage
+
+The main `.github/workflows/test.yml` workflow runs this contract on pushes and
+pull requests. The standalone `.github/workflows/clean-host-contract.yml`
+workflow remains available for manual runs and weekly scheduled drift checks.
+CI uses the default fake-host path and does not spend model quota.
+
+## Failure Triage
+
+- Wheel build failure: inspect package metadata and `pyproject.toml`.
+- Install failure: inspect dependency constraints and `pip check` output.
+- `ctx-init` failure: inspect packaged entrypoints and hook module paths.
+- Fake Claude hook-smoke failure: inspect generated `settings.json`, packaged
+  hook module paths, and whether PostToolUse/Stop hook schemas changed.
+- Live Claude gate failure: inspect explicit auth env/provider credentials,
+  `claude auth status`, the budget cap, the injected sentinel hook entries, and
+  `live-claude-hooks.jsonl` under the temp root.
+- `ctx-scan-repo` failure: inspect installed flat-module entrypoints and
+  resolver imports.
+- `ctx run` or `ctx resume` failure: inspect LiteLLM provider import behavior,
+  session store paths, and CLI metadata replay.
+- Tool denial failure: inspect `--allow-tool`/`--deny-tool` policy handling in
+  `src/ctx/cli/run.py` and `src/ctx/adapters/generic/loop.py`.

docs/huggingface-publish.md

@@ -1,67 +1,79 @@
-# Hugging Face Publish
-
-ctx publishes the GitHub repository as the public Hugging Face dataset
-[`Stevesolun/ctx`](https://huggingface.co/datasets/Stevesolun/ctx). The
-dataset is a clean `git ls-files` snapshot, including the shipped graph
-tarball and catalog artifacts, not local review reports or ignored caches.
-
-## What gets uploaded
-
-- Tracked source, docs, tests, and packaging files.
-- `graph/wiki-graph.tar.gz`.
-- `graph/wiki-graph-runtime.tar.gz`.
-- The compressed skill index under `graph/`.
-- Tracked graph visualizations under `graph/`.
-
-Ignored local reports, review notes, raw ingest caches, coverage files,
-`site/`, and `.pytest_cache/` are not uploaded because they are not tracked
-by git.
-
-## Publish command
-
-Use the repository sync script. It exports only tracked files, adds the
-Hugging Face repo-card frontmatter to the uploaded `README.md`, and refuses to
-publish if the full wiki tarball, runtime wiki tarball, or compressed skill
-index is missing, too small, or still a Git LFS
-pointer.
-
-The script prefers Hugging Face's resumable large-folder uploader when the
-remote already has no stale paths. If the remote contains files that are not in
-the current git snapshot, the script falls back to a single clean replacement
-commit so deleted local files cannot survive remotely.
-
-Do not paste the token into a command line. Prompt for it, set it only for the
-current process, and clear it after the upload.
-
-```powershell
-python -m pip install --upgrade huggingface_hub
-git lfs install
-git lfs pull --include="graph/wiki-graph.tar.gz,graph/wiki-graph-runtime.tar.gz"
-
-$secureToken = Read-Host "HF write token" -AsSecureString
-$tokenPtr = [Runtime.InteropServices.Marshal]::SecureStringToBSTR($secureToken)
-try {
-  $env:HF_TOKEN = [Runtime.InteropServices.Marshal]::PtrToStringBSTR($tokenPtr)
-  python scripts/sync_huggingface.py --repo . --repo-id Stevesolun/ctx --repo-type dataset
-} finally {
-  if ($tokenPtr -ne [IntPtr]::Zero) {
-    [Runtime.InteropServices.Marshal]::ZeroFreeBSTR($tokenPtr)
-  }
-  Remove-Item Env:\HF_TOKEN -ErrorAction SilentlyContinue
-}
-```
-
-## Verify
-
-```powershell
-@'
-from huggingface_hub import HfApi
-
-api = HfApi()
-info = api.repo_info(repo_id="Stevesolun/ctx", repo_type="dataset")
-print(info.id, info.sha)
-'@ | python -
-```
-
-The dataset page should show the MIT license and the tags from the metadata
-wrapper.
+# Hugging Face Publish
+
+ctx publishes the GitHub repository as the public Hugging Face dataset
+[`Stevesolun/ctx`](https://huggingface.co/datasets/Stevesolun/ctx). The
+dataset is a clean `git ls-files` snapshot, including the shipped graph
+tarball and catalog artifacts, not local review reports or ignored caches.
+
+## What gets uploaded
+
+- Tracked source, docs, tests, and packaging files.
+- `graph/wiki-graph.tar.gz`.
+- `graph/wiki-graph-runtime.tar.gz`.
+- The compressed skill index under `graph/`.
+- Tracked graph visualizations under `graph/`.
+
+Ignored local reports, review notes, raw ingest caches, coverage files,
+`site/`, and `.pytest_cache/` are not uploaded because they are not tracked
+by git.
+
+## Automatic publish
+
+Every push to `main` runs `.github/workflows/huggingface-sync.yml`. The job
+checks out source without spending Git LFS bandwidth, hydrates the required
+graph artifacts from the latest GitHub release assets, installs the sync
+dependencies, and calls `scripts/sync_huggingface.py`. It publishes only when
+the repository secret `HF_TOKEN` is configured; otherwise it exits successfully
+with a notice so public forks and dry repos do not fail.
+
+The sync script is still the contract: it exports the tracked git snapshot,
+adds Hugging Face repo-card metadata, validates README/docs stats, verifies the
+graph artifacts are hydrated rather than LFS pointers, and refuses to publish
+stale or corrupt artifacts.
+
+## Manual publish
+
+Use the repository sync script. It exports tracked files plus the validated
+local graph artifacts, adds the Hugging Face repo-card frontmatter to the
+uploaded `README.md`, and refuses to publish if the full wiki tarball, runtime
+wiki tarball, or compressed skill index is missing, too small, or still a Git
+LFS pointer.
+
+The script prefers Hugging Face's resumable large-folder uploader when the
+remote already has no stale paths. If the remote contains files that are not in
+the current git snapshot, the script falls back to a single clean replacement
+commit so deleted local files cannot survive remotely.
+
+Do not paste the token into a command line. Prompt for it, set it only for the
+current process, and clear it after the upload.
+
+```powershell
+python -m pip install --upgrade huggingface_hub
+
+$secureToken = Read-Host "HF write token" -AsSecureString
+$tokenPtr = [Runtime.InteropServices.Marshal]::SecureStringToBSTR($secureToken)
+try {
+  $env:HF_TOKEN = [Runtime.InteropServices.Marshal]::PtrToStringBSTR($tokenPtr)
+  python scripts/sync_huggingface.py --repo . --repo-id Stevesolun/ctx --repo-type dataset
+} finally {
+  if ($tokenPtr -ne [IntPtr]::Zero) {
+    [Runtime.InteropServices.Marshal]::ZeroFreeBSTR($tokenPtr)
+  }
+  Remove-Item Env:\HF_TOKEN -ErrorAction SilentlyContinue
+}
+```
+
+## Verify
+
+```powershell
+@'
+from huggingface_hub import HfApi
+
+api = HfApi()
+info = api.repo_info(repo_id="Stevesolun/ctx", repo_type="dataset")
+print(info.id, info.sha)
+'@ | python -
+```
+
+The dataset page should show the MIT license and the tags from the metadata
+wrapper.

docs/index.md

@@ -1,232 +1,232 @@
----
-hide:
-  - navigation
----
-
-# ctx — Skill, Agent, MCP & Harness Recommendations
-
-[![Repo views](https://hits.sh/github.com/stevesolun/ctx.svg?label=repo%20views)](https://hits.sh/github.com/stevesolun/ctx/)
-
-Watches what you develop, walks a knowledge graph of **91,464 skill pages, 467 agents, 10,790 MCP servers, and 207 cataloged harnesses**, and recommends the
-right execution bundle on the fly. The live execution bundle is skills,
-agents, and MCP servers only; custom/API/local model users get a separate
-harness recommendation based on model choice and task goal. You decide
-what to load, install, or adopt. Powered by a Karpathy LLM wiki with persistent
-memory that gets smarter every session.
-
-!!! tip "Install"
-
-    ```bash
-    pip install claude-ctx
-    ctx-init --graph --model-mode skip
-    ```
-
-    Optional extras: `pip install "claude-ctx[embeddings]"` for the
-    semantic backend, `pip install "claude-ctx[harness]"` for local/API
-    model harness runs, `pip install "claude-ctx[dev]"` for the
-    pytest/mypy/ruff toolchain. After install the `ctx-scan-repo`,
-    `ctx-skill-quality`, `ctx-skill-health`, and `ctx-toolbox` console
-    scripts are on PATH. `ctx-init --graph` installs the fast pre-built
-    runtime graph that powers recommendations and harness dry-runs; source checkouts use
-    `graph/wiki-graph-runtime.tar.gz`, while pip installs download the
-    matching GitHub release asset. Use
-    `ctx-init --graph --graph-install-mode full` when you want the full
-    markdown LLM-wiki expanded locally.
-
-    Custom-model users can run
-    `ctx-init --model-mode custom --model <provider/model> --goal "<task>"`
-    to record the model profile and surface harness recommendations.
-
-!!! tip "Before pushing"
-
-    ```bash
-    python scripts/ci_preflight.py --profile pr
-    ```
-
-    The preflight uses the same changed-file classifier as GitHub Actions and
-    runs the matching local gates before you open a PR: stats, ruff, mypy, pip
-    check, unit coverage, canaries, package build, twine, docs, graph
-    validation, browser, and similarity checks as needed. Use `--profile full`
-    before release work to force the source/package gates even for docs-only or
-    graph-only changes.
-
-## Why this exists
-
-Claude Code skills, agents, MCP servers, and model harness profiles are
-powerful, but at scale they become unmanageable:
-
-- **Discovery problem** — with 91K+ skill pages, 460+ agents, 10,000+
-  MCP servers, and 207 harnesses, how do you know which
-  ones exist and which are relevant to your current project?
-- **Context budget** — loading every installable entity wastes tokens and
-  degrades quality. You need exactly the right skills, agents, and MCP
-  servers per session, plus a harness recommendation only when you choose
-  a custom/API/local model path.
-- **Hidden connections** — a FastAPI skill is useful, but you also need
-  the Pydantic skill, the async Python patterns skill, and the Docker
-  skill, plus possibly a matching MCP server. If you are not using Claude
-  Code, ctx separately suggests the model harness most likely to fit your
-  goal.
-  Nobody tells you that.
-- **Entity rot** — skills, agents, MCP servers, and harness records you
-  added months ago and never used are cluttering your context. Stale ones
-  should be flagged and archived.
-
-ctx solves all of these by treating your ctx inventory as a **knowledge
-graph with persistent memory**, not a flat directory.
-
-## What this is
-
-ctx is not a collection of scripts. It is an agent with persistent memory
-and a knowledge graph.
-
-The core idea comes from Andrej Karpathy's LLM-wiki pattern: instead of
-re-loading everything from scratch each session, an LLM maintains a wiki
-it can read, write, and query. The wiki becomes the agent's long-term
-memory.
-
-ctx applies that pattern to entity management — and extends it with
-graph-based discovery:
-
-- A Karpathy 3-layer wiki at `~/.claude/skill-wiki/` is the single source
-  of truth.
-- **102,928 graph nodes** for the shipped skill/agent/MCP/harness
-  inventory, including 91,464 skill pages
-  and 207 harness pages under `entities/harnesses/`.
-  Each page tracks tags, status, provenance, and usage where it applies.
-- A **knowledge graph** (102,928 nodes, 2,913,960 edges) built from a
-  13,463-node core plus 89,465 body-backed skill nodes.
-  The graph has 52 Louvain communities and blends semantic cosine,
-  tag overlap, and slug-token overlap; 89,465 skill bodies are
-  shipped as installable `SKILL.md` files. Entries over the configured line
-  threshold are converted to gated micro-skill orchestrators. Full source
-  bodies were used for semantic graphing before packaging; `SKILL.md.original`
-  backups are not shipped in the tarball.
-- **52 Louvain communities** group related entities into named
-  communities (e.g., *AI + Devops + Frontend*, *Python + API*).
-- PostToolUse and Stop hooks update the wiki automatically during each
-  Claude Code session.
-- Hydrated skills over 180 lines are converted to gated micro-skill
-  pipelines so the router can load them incrementally.
-- At session start, the skill-router scans your project and
-  **recommends** the best-matching skills, agents, and MCP servers.
-- Mid-session, the context monitor watches every tool call, detects new
-  stack signals, walks the graph, and **recommends** relevant skills,
-  agents, and MCP servers in real time — **nothing loads or
-  installs without your approval**.
-- During custom/API/local model onboarding, `ctx-init` and
-  `ctx-harness-install` use the same graph to recommend harnesses
-  above the configured harness match floor.
-
-The result: you always know what skills, agents, and MCP servers are available
-for your current task, and which harness fits when you choose your own model.
-The graph reveals hidden connections. The wiki learns from your usage. Stale
-ones are flagged. New ones self-ingest.
-
-## Explore the docs
-
-<div class="grid cards" markdown>
-
--   **Knowledge graph**
-
-    ---
-
-    102,928 shipped graph nodes: 13,463 curated skill/agent/MCP/harness nodes plus 89,465 body-backed skill nodes. The graph has
-    2,913,960 weighted edges and 52 Louvain communities.
-    Ships pre-built in `graph/wiki-graph.tar.gz` and powers the
-    graph-aware recommendations + the pre-ship `ctx-dedup-check` gate.
-
-    [:octicons-arrow-right-24: Knowledge graph](knowledge-graph.md)
-
--   **Entity onboarding**
-
-    ---
-
-    Step-by-step commands for adding a skill, agent, MCP server, or
-    harness to the wiki and graph. Includes the `text-to-cad` harness
-    pattern for custom-model users.
-
-    [:octicons-arrow-right-24: Entity onboarding](entity-onboarding.md)
-
--   **Dashboard**
-
-    ---
-
-    `ctx-monitor serve` opens a local HTTP dashboard with live graph,
-    skill grades + four-signal scores, session timelines, one-click
-    load/unload for skills, agents, and MCP servers, plus harness wiki
-    and graph browsing. It is served by stdlib `http.server` and renders
-    repo docs with MkDocs-compatible Markdown extensions.
-
-    [:octicons-arrow-right-24: Dashboard reference](dashboard.md)
-
--   **Toolbox**
-
-    ---
-
-    Curated councils of skills and agents that fire at session-start,
-    file-save, pre-commit, and session-end. Blocks `git commit` on
-    HIGH/CRITICAL findings. Five starter toolboxes ship out of the box.
-
-    [:octicons-arrow-right-24: Toolbox overview](toolbox/index.md) ·
-    [Starter toolboxes](toolbox/starters.md) ·
-    [Verdicts & guardrails](toolbox/verdicts.md)
-
--   **Skill router**
-
-    ---
-
-    Scans the active repo, detects the stack from file signatures, walks
-    the stack matrix, loads exactly the skills that apply, and can
-    recommend supporting agents and MCP servers.
-
-    [:octicons-arrow-right-24: Router overview](skill-router/index.md) ·
-    [Stack signatures](stack-signatures.md) ·
-    [Skill-stack matrix](skill-stack-matrix.md)
-
--   **Health & quality**
-
-    ---
-
-    Structural health checks (missing frontmatter, orphan manifest
-    entries, line-count drift) plus the four-signal quality score
-    (telemetry · intake · graph · routing) that grades every skill
-    A/B/C/D/F.
-
-    [:octicons-arrow-right-24: Skill health](skills-health.md) ·
-    [Memory anchoring](memory-anchor.md) ·
-    [Lifecycle dashboard](skill-lifecycle-and-dashboard.md)
-
--   **Releases**
-
-    ---
-
-    **v1.0.11** — MIT, CI-matrixed (Ubuntu 3.12 plus Windows/macOS 3.11/3.12),
-    3,915 tests collected. Ships console scripts including `ctx-init`,
-    `ctx-monitor` (local dashboard with graph + wiki + load/unload for
-    skills, agents, and MCP servers, plus Harness Setup for user-owned LLMs),
-    `ctx-incremental-attach`, `ctx-incremental-shadow`, `ctx-dedup-check`
-    (pre-ship near-duplicate gate), and
-    `ctx-tag-backfill` (entity hygiene), plus a fast runtime graph artifact
-    and the full ~439 MiB wiki tarball with **102,928 nodes / 2,913,960 edges / 52 Louvain communities**.
-
-    [:octicons-arrow-right-24: CHANGELOG](https://github.com/stevesolun/ctx/blob/main/CHANGELOG.md) ·
-    [Repository](https://github.com/stevesolun/ctx)
-
-</div>
-
-## Principles
-
-- **Single source of truth.** The wiki and graph drive Claude Code
-  recommendations, custom-model harness recommendations, dashboard views,
-  and entity update reviews.
-- **Explicit approval.** ctx can recommend, review, install, update, unload,
-  or uninstall, but it does not mutate live skills, agents, MCP servers, or
-  harness installs without a command or approval path.
-- **Configurable gates.** Recommendation floors, semantic edge thresholds,
-  micro-skill line limits, and harness match floors live in config so teams
-  can tune behavior without forking the code.
-- **Evidence over opinion.** Suggestions cite real usage data plus
-  knowledge-graph edges. No black-box prompts.
-- **Token discipline.** Every council run honors `max_tokens` /
-  `max_seconds` budgets.
+---
+hide:
+  - navigation
+---
+
+# ctx — Skill, Agent, MCP & Harness Recommendations
+
+[![Repo views](https://hits.sh/github.com/stevesolun/ctx.svg?label=repo%20views)](https://hits.sh/github.com/stevesolun/ctx/)
+
+Watches what you develop, walks a knowledge graph of **91,464 skill pages, 467 agents, 10,790 MCP servers, and 207 cataloged harnesses**, and recommends the
+right execution bundle on the fly. The live execution bundle is skills,
+agents, and MCP servers only; custom/API/local model users get a separate
+harness recommendation based on model choice and task goal. You decide
+what to load, install, or adopt. Powered by a Karpathy LLM wiki with persistent
+memory that gets smarter every session.
+
+!!! tip "Install"
+
+    ```bash
+    pip install claude-ctx
+    ctx-init --graph --model-mode skip
+    ```
+
+    Optional extras: `pip install "claude-ctx[embeddings]"` for the
+    semantic backend, `pip install "claude-ctx[harness]"` for local/API
+    model harness runs, `pip install "claude-ctx[dev]"` for the
+    pytest/mypy/ruff toolchain. After install the `ctx-scan-repo`,
+    `ctx-skill-quality`, `ctx-skill-health`, and `ctx-toolbox` console
+    scripts are on PATH. `ctx-init --graph` installs the fast pre-built
+    runtime graph that powers recommendations and harness dry-runs; source checkouts use
+    `graph/wiki-graph-runtime.tar.gz`, while pip installs download the
+    matching GitHub release asset. Use
+    `ctx-init --graph --graph-install-mode full` when you want the full
+    markdown LLM-wiki expanded locally.
+
+    Custom-model users can run
+    `ctx-init --model-mode custom --model <provider/model> --goal "<task>"`
+    to record the model profile and surface harness recommendations.
+
+!!! tip "Before pushing"
+
+    ```bash
+    python scripts/ci_preflight.py --profile pr
+    ```
+
+    The preflight uses the same changed-file classifier as GitHub Actions and
+    runs the matching local gates before you open a PR: stats, ruff, mypy, pip
+    check, unit coverage, canaries, package build, twine, docs, graph
+    validation, browser, and similarity checks as needed. Use `--profile full`
+    before release work to force the source/package gates even for docs-only or
+    graph-only changes.
+
+## Why this exists
+
+Claude Code skills, agents, MCP servers, and model harness profiles are
+powerful, but at scale they become unmanageable:
+
+- **Discovery problem** — with 91K+ skill pages, 460+ agents, 10,000+
+  MCP servers, and 207 harnesses, how do you know which
+  ones exist and which are relevant to your current project?
+- **Context budget** — loading every installable entity wastes tokens and
+  degrades quality. You need exactly the right skills, agents, and MCP
+  servers per session, plus a harness recommendation only when you choose
+  a custom/API/local model path.
+- **Hidden connections** — a FastAPI skill is useful, but you also need
+  the Pydantic skill, the async Python patterns skill, and the Docker
+  skill, plus possibly a matching MCP server. If you are not using Claude
+  Code, ctx separately suggests the model harness most likely to fit your
+  goal.
+  Nobody tells you that.
+- **Entity rot** — skills, agents, MCP servers, and harness records you
+  added months ago and never used are cluttering your context. Stale ones
+  should be flagged and archived.
+
+ctx solves all of these by treating your ctx inventory as a **knowledge
+graph with persistent memory**, not a flat directory.
+
+## What this is
+
+ctx is not a collection of scripts. It is an agent with persistent memory
+and a knowledge graph.
+
+The core idea comes from Andrej Karpathy's LLM-wiki pattern: instead of
+re-loading everything from scratch each session, an LLM maintains a wiki
+it can read, write, and query. The wiki becomes the agent's long-term
+memory.
+
+ctx applies that pattern to entity management — and extends it with
+graph-based discovery:
+
+- A Karpathy 3-layer wiki at `~/.claude/skill-wiki/` is the single source
+  of truth.
+- **102,928 graph nodes** for the shipped skill/agent/MCP/harness
+  inventory, including 91,464 skill pages
+  and 207 harness pages under `entities/harnesses/`.
+  Each page tracks tags, status, provenance, and usage where it applies.
+- A **knowledge graph** (102,928 nodes, 2,913,960 edges) built from a
+  13,463-node core plus 89,465 body-backed skill nodes.
+  The graph has 52 Louvain communities and blends semantic cosine,
+  tag overlap, and slug-token overlap; 89,465 skill bodies are
+  shipped as installable `SKILL.md` files. Entries over the configured line
+  threshold are converted to gated micro-skill orchestrators. Full source
+  bodies were used for semantic graphing before packaging; `SKILL.md.original`
+  backups are not shipped in the tarball.
+- **52 Louvain communities** group related entities into named
+  communities (e.g., *AI + Devops + Frontend*, *Python + API*).
+- PostToolUse and Stop hooks update the wiki automatically during each
+  Claude Code session.
+- Hydrated skills over 180 lines are converted to gated micro-skill
+  pipelines so the router can load them incrementally.
+- At session start, the skill-router scans your project and
+  **recommends** the best-matching skills, agents, and MCP servers.
+- Mid-session, the context monitor watches every tool call, detects new
+  stack signals, walks the graph, and **recommends** relevant skills,
+  agents, and MCP servers in real time — **nothing loads or
+  installs without your approval**.
+- During custom/API/local model onboarding, `ctx-init` and
+  `ctx-harness-install` use the same graph to recommend harnesses
+  above the configured harness match floor.
+
+The result: you always know what skills, agents, and MCP servers are available
+for your current task, and which harness fits when you choose your own model.
+The graph reveals hidden connections. The wiki learns from your usage. Stale
+ones are flagged. New ones self-ingest.
+
+## Explore the docs
+
+<div class="grid cards" markdown>
+
+-   **Knowledge graph**
+
+    ---
+
+    102,928 shipped graph nodes: 13,463 curated skill/agent/MCP/harness nodes plus 89,465 body-backed skill nodes. The graph has
+    2,913,960 weighted edges and 52 Louvain communities.
+    Ships pre-built in `graph/wiki-graph.tar.gz` and powers the
+    graph-aware recommendations + the pre-ship `ctx-dedup-check` gate.
+
+    [:octicons-arrow-right-24: Knowledge graph](knowledge-graph.md)
+
+-   **Entity onboarding**
+
+    ---
+
+    Step-by-step commands for adding a skill, agent, MCP server, or
+    harness to the wiki and graph. Includes the `text-to-cad` harness
+    pattern for custom-model users.
+
+    [:octicons-arrow-right-24: Entity onboarding](entity-onboarding.md)
+
+-   **Dashboard**
+
+    ---
+
+    `ctx-monitor serve` opens a local HTTP dashboard with live graph,
+    skill grades + four-signal scores, session timelines, one-click
+    load/unload for skills, agents, and MCP servers, plus harness wiki
+    and graph browsing. It is served by stdlib `http.server` and renders
+    repo docs with MkDocs-compatible Markdown extensions.
+
+    [:octicons-arrow-right-24: Dashboard reference](dashboard.md)
+
+-   **Toolbox**
+
+    ---
+
+    Curated councils of skills and agents that fire at session-start,
+    file-save, pre-commit, and session-end. Blocks `git commit` on
+    HIGH/CRITICAL findings. Five starter toolboxes ship out of the box.
+
+    [:octicons-arrow-right-24: Toolbox overview](toolbox/index.md) ·
+    [Starter toolboxes](toolbox/starters.md) ·
+    [Verdicts & guardrails](toolbox/verdicts.md)
+
+-   **Skill router**
+
+    ---
+
+    Scans the active repo, detects the stack from file signatures, walks
+    the stack matrix, loads exactly the skills that apply, and can
+    recommend supporting agents and MCP servers.
+
+    [:octicons-arrow-right-24: Router overview](skill-router/index.md) ·
+    [Stack signatures](stack-signatures.md) ·
+    [Skill-stack matrix](skill-stack-matrix.md)
+
+-   **Health & quality**
+
+    ---
+
+    Structural health checks (missing frontmatter, orphan manifest
+    entries, line-count drift) plus the four-signal quality score
+    (telemetry · intake · graph · routing) that grades every skill
+    A/B/C/D/F.
+
+    [:octicons-arrow-right-24: Skill health](skills-health.md) ·
+    [Memory anchoring](memory-anchor.md) ·
+    [Lifecycle dashboard](skill-lifecycle-and-dashboard.md)
+
+-   **Releases**
+
+    ---
+
+    **v1.0.11** — MIT, CI-matrixed (Ubuntu 3.12 plus Windows/macOS 3.11/3.12),
+    3,923 tests collected. Ships console scripts including `ctx-init`,
+    `ctx-monitor` (local dashboard with graph + wiki + load/unload for
+    skills, agents, and MCP servers, plus Harness Setup for user-owned LLMs),
+    `ctx-incremental-attach`, `ctx-incremental-shadow`, `ctx-dedup-check`
+    (pre-ship near-duplicate gate), and
+    `ctx-tag-backfill` (entity hygiene), plus a fast runtime graph artifact
+    and the full ~439 MiB wiki tarball with **102,928 nodes / 2,913,960 edges / 52 Louvain communities**.
+
+    [:octicons-arrow-right-24: CHANGELOG](https://github.com/stevesolun/ctx/blob/main/CHANGELOG.md) ·
+    [Repository](https://github.com/stevesolun/ctx)
+
+</div>
+
+## Principles
+
+- **Single source of truth.** The wiki and graph drive Claude Code
+  recommendations, custom-model harness recommendations, dashboard views,
+  and entity update reviews.
+- **Explicit approval.** ctx can recommend, review, install, update, unload,
+  or uninstall, but it does not mutate live skills, agents, MCP servers, or
+  harness installs without a command or approval path.
+- **Configurable gates.** Recommendation floors, semantic edge thresholds,
+  micro-skill line limits, and harness match floors live in config so teams
+  can tune behavior without forking the code.
+- **Evidence over opinion.** Suggestions cite real usage data plus
+  knowledge-graph edges. No black-box prompts.
+- **Token discipline.** Every council run honors `max_tokens` /
+  `max_seconds` budgets.

docs/knowledge-graph.md

@@ -1,359 +1,359 @@
-# Knowledge graph
-
-A pre-built weighted graph of skills, agents, MCP servers, and
-harnesses in the ctx ecosystem, shipped as `graph/wiki-graph.tar.gz`.
-The on-disk JSON and `resolve_graph` Python API are harness-aware, including
-plain-slug graph walks from `harness:<slug>` nodes. `ctx-monitor`
-exposes skill/agent/MCP/harness wiki and graph views. Harness installation,
-update, and uninstall are handled by `ctx-harness-install`; dashboard
-load/unload POSTs deliberately reject harnesses and return the dry-run CLI
-command to use instead. Quality scoring is exposed for sidecar-backed skills,
-agents, and MCP servers.
-
-## What's in it
-
-Authoritative numbers from the shipped tarball. The curated-core snapshot
-is **13,463 nodes** (1,999 curated skills + 467 agents + 10,790 MCP servers + 207 harnesses). Harness pages under `entities/harnesses/` are ingested into
-local rebuilds and the separate harness recommendation path. The
-tarball also carries **91,464 skill pages**; **89,465**
-skill bodies are hydrated as installable `SKILL.md` files under
-`converted/`; the **28,612** entries over the configured line
-limit were converted to gated micro-skill orchestrators. Full original bodies
-are used during graph rebuilds for semantic similarity, but
-`SKILL.md.original` backups, transient `.lock` files, and `.ctx/` queue state
-are omitted from the shipped tarball.
-
-| | Count |
-|---|---:|
-| Total nodes | **102,928** |
-| Curated core nodes | **13,463** (1,999 skills + 467 agents + 10,790 MCP servers + 207 harnesses) |
-| Body-backed skill nodes | **89,465** hydrated installable skill entries |
-| Total edges | **2,913,960** |
-| Hydrated skill incident edges | **2,605,721** |
-| Hydrated skill semantic incident edges | **1,500,648** |
-| Communities | **52** (Louvain) |
-| Edge sources (overlap-deduped) | semantic 1,683,193 - tag 897,784 - token 433,245 |
-| Cross-type edges (skill <-> agent) | ~66,799 |
-| Cross-type edges (skill <-> MCP) | ~41,521 |
-| Cross-type edges (agent <-> MCP) | ~229 |
-| Harness edges | **6,576** |
-| Shipped skill index | **89,465** observed body-backed skill entries |
-
-## Install
-
-Use `ctx-init --graph` to install the fast runtime graph. Source checkouts use
-`graph/wiki-graph-runtime.tar.gz`; pip installs download the matching GitHub
-release asset for the installed package version. This installs
-`graphify-out/*`, the skill index used by recommendations, and
-the harness pages used by `ctx-harness-install`:
-
-```bash
-ctx-init --graph
-```
-
-To expand every shipped skill/agent/MCP entity page, harness page,
-skill page, concept page, converted micro-skill pipeline,
-and Obsidian vault metadata, request the full wiki artifact explicitly:
-
-```bash
-ctx-init --graph --graph-install-mode full
-```
-
-Manual extraction is still supported for offline/source installs. Extract the
-full tarball into your `~/.claude/skill-wiki/` when you want local markdown
-wiki browsing:
-
-```bash
-mkdir -p ~/.claude/skill-wiki
-tar xzf graph/wiki-graph.tar.gz -C ~/.claude/skill-wiki/
-```
-
-On Windows PowerShell, create the target and use the built-in `tar.exe`
-without `--force-local`:
-
-```powershell
-New-Item -ItemType Directory -Force "$env:USERPROFILE\.claude\skill-wiki" | Out-Null
-tar -xzf graph\wiki-graph.tar.gz -C "$env:USERPROFILE\.claude\skill-wiki"
-```
-
-The extracted tree also opens directly as an Obsidian vault — the
-`.obsidian/` config ships inside the tarball — so you can use
-Obsidian's native graph view if you prefer it to the web dashboard.
-
-## How edges are built
-
-Edges are built and explained by the `ctx-wiki-graphify` console script
-(`ctx.core.wiki.wiki_graphify`). A pair must first have at least one base
-signal:
-
-1. **Semantic cosine** — when the embedding backend is available, entity
-   text is embedded and semantic neighbors above the configured build floor
-   contribute weighted edges.
-2. **Explicit frontmatter tags** — each entity page's YAML `tags:`
-   list contributes edges between every pair of entities that share
-   a tag. Popular tags capped at 500 nodes to avoid noise-floor
-   "everything connects to everything" mega-buckets like `typescript`
-   or `frontend`.
-3. **Slug-token pseudo-tags** — each hyphenated slug contributes its
-   tokens as implicit tags. `fastapi-pro` contributes `fastapi`;
-   `python-patterns` contributes `python` and `patterns`. A stop-word
-   filter drops generic tokens like `skill`, `agent`, `pro`, `expert`,
-   `core` so they don't over-connect the graph.
-4. **Source overlap** — pages with the same high-specificity source URL,
-   repository URL, homepage, detail URL, or package URL can connect even
-   when their tags differ. Dense source buckets are skipped.
-5. **Direct wikilinks** — explicit entity links such as
-   `[[entities/agents/code-reviewer]]` create a direct graph edge.
-
-Edge `weight` is the final blended strength. Semantic, tag, and token
-weights form the base blend from `config.json`; source overlap and direct
-links add configured boosts. Existing edges can also receive explainable
-ranking boosts from Adamic-Adar shared-neighbor structure, type affinity,
-usage telemetry, and quality scores. Those boost-only signals do not create
-edges by themselves. The shipped default `graph.min_edge_weight` is `0.03`;
-calibration against the 2026-05 shipped graph showed this is the highest
-floor with zero edge loss, while `0.05` would remove roughly 29.7% of edges.
-
-Edge metadata keeps the ingredients explainable: `semantic_sim`,
-`shared_tags`, `shared_tokens`, `shared_sources`, `direct_link`,
-`adamic_adar`, `type_affinity`, `usage_score`, `quality_score`,
-`edge_reasons`, and `score_components`. Hydrated skill records use their
-full source bodies during graph rebuilds, so long converted entries keep
-full-body similarity even though the shipped installable `SKILL.md` files are
-short gated loaders. The raw `SKILL.md.original` backups are build inputs, not
-tarball members.
-
-## Communities
-
-After edges are built, `wiki_graphify` runs NetworkX's Louvain
-community detection (`resolution=1.2`, `seed=42` for determinism).
-The result is **52 communities** ranging from single-member isolated
-specialists to several thousand members in broad clusters like
-`Community + Official + AI`. Each community also gets an auto-generated
-`concepts/<community>.md` wiki page summarizing its members and top
-shared tags.
-
-The legacy CNM ("greedy modularity") algorithm is still available
-behind `CTX_GRAPH_COMMUNITY=cnm` — it's deterministic but O(n²) on
-dense graphs and hangs on the live 13K-node dataset (~50min run was
-killed on 2026-04-27 inside the priority-queue siftup). Louvain is
-the default because it finishes in seconds and produces equivalent
-quality clusters for the recommendation use case.
-
-## Querying the graph
-
-### Via the dashboard
-
-```bash
-ctx-monitor serve              # http://127.0.0.1:8765
-```
-
-Then open `/graph?slug=<entity-slug>&type=<entity-type>` for a
-cytoscape neighborhood view, or
-`/api/graph/<slug>.json?type=<entity-type>&hops=1&limit=40` for the
-dashboard-shaped JSON. The `type` query is optional for unique slugs and
-recommended for duplicate slugs such as `langgraph`. See the
-[dashboard reference](dashboard.md) for the full route catalogue.
-
-### Via Python
-
-```python
-import json
-from pathlib import Path
-from networkx.readwrite import node_link_graph
-
-raw = json.loads(
-    Path("~/.claude/skill-wiki/graphify-out/graph.json").expanduser().read_text()
-)
-edges_key = "links" if "links" in raw else "edges"
-G = node_link_graph(raw, edges=edges_key)
-
-# 102,928 nodes, 2,913,960 edges
-print(G.number_of_nodes(), G.number_of_edges())
-
-# Find entities related to 'fastapi-pro' by edge weight
-seed = "skill:fastapi-pro"
-neighbors = sorted(
-    G.neighbors(seed),
-    key=lambda n: G[seed][n]["weight"],
-    reverse=True,
-)[:10]
-for n in neighbors:
-    shared = G[seed][n].get("shared_tags", [])
-    print(f"  w={G[seed][n]['weight']:>2}  {G.nodes[n]['label']:<40}  {shared[:3]}")
-```
-
-The node-link JSON schema's edges key is auto-detected (legacy
-NetworkX 2.x used `"links"`; current versions default to `"edges"`).
-The helper `resolve_graph.load_graph()` does this for you.
-
-### Via recommendation paths
-
-The graph backs two recommendation paths:
-
-- Execution recommendation surfaces (`ctx.recommend_bundle`, MCP
-  `ctx__recommend_bundle`, generic harness tools, Claude Code hook
-  suggestions, and repo-scan advisory output) share
-  `ctx.core.resolve.recommendations.recommend_by_tags` for skills,
-  agents, and MCP servers. That engine ranks candidates by
-  slug-token matches, tag overlap, graph degree, and semantic-cache
-  signals when available. Imported skill results are normal `skill` nodes with
-  detail URLs, install commands, duplicate
-  hints, gated micro-skill loaders when over the line threshold, and
-  quality/security metadata. If an older
-  extracted wiki has the skill index JSON but no graph nodes for
-  those records, the same recommender falls back to the index file.
-- Harness recommendations are a separate path for custom/API/local
-  model onboarding (`ctx-init --model-mode custom ...`) and
-  `ctx-harness-install`. They use the same graph filtered to
-  `harness` nodes and the higher harness match floor from `config.json`.
-- Repository scans still start from stack detections, then turn that profile
-  into the same tag/query bundle used by the execution recommender. If a
-  shipped graph is unavailable, scan output falls back to the legacy installed
-  skill resolver so a plain profile scan remains useful. Harnesses are
-  intentionally not emitted from repo scans or Claude Code hook bundles.
-
-This split is intentional: execution surfaces need identical ranking and a
-small top-K, while harness choice changes the model runtime itself and belongs
-in an explicit onboarding/install flow.
-
-### LLM-wiki design references
-
-ctx follows Karpathy's LLM-wiki pattern. We also reviewed
-[`nashsu/llm_wiki`](https://github.com/nashsu/llm_wiki) as a design reference
-for source traceability, persistent ingest queues, graph insights, and
-budgeted token/vector/graph retrieval. That repository is GPLv3, while ctx is
-MIT, so ctx can use those ideas as product inspiration but must not copy or
-vendor its code or assets.
-
-## Rebuilding
-
-After you add a skill, agent, MCP server, or harness entity page:
-
-```bash
-ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1
-```
-
-The `entity-upsert` worker path validates the queued page hash, updates the
-wiki index, and, when a persisted semantic vector index exists, runs a
-best-effort ANN attach into `graphify-out/entity-overlays.jsonl`. That overlay
-lets the runtime resolver connect a new or updated entity to existing graph
-neighbors without recomputing global all-pairs similarity. The worker still
-queues the normal incremental `graph-export` job, and the entity markdown page
-remains the source of truth.
-
-For manual review or debugging:
-
-```bash
-ctx-incremental-attach calibrate \
-  --graph ~/.claude/skill-wiki/graphify-out/graph.json
-
-ctx-incremental-attach attach \
-  --index-dir ~/.claude/skill-wiki/.embedding-cache/graph/vector-index \
-  --overlay ~/.claude/skill-wiki/graphify-out/entity-overlays.jsonl \
-  --node-id skill:fastapi-review \
-  --type skill \
-  --label fastapi-review \
-  --text-file ~/.claude/skill-wiki/entities/skills/fastapi-review.md \
-  --dry-run
-```
-
-Shadow-gate a persisted index before trusting a new ANN backend, changed
-thresholds, or a large attach workflow:
-
-```bash
-ctx-incremental-shadow \
-  --index-dir ~/.claude/skill-wiki/.embedding-cache/graph/vector-index \
-  --graph ~/.claude/skill-wiki/graphify-out/graph.json \
-  --sample-size 100 \
-  --min-overlap 0.85
-```
-
-The shadow command pretends sampled existing nodes are new, compares the
-incremental attach result to batch graph semantic neighbors, and reports
-precision, recall, top-5/top-10/top-20 agreement, score deltas, and bad
-examples. A failing gate means either tune thresholds or use a full graph
-rebuild before shipping.
-
-If the vector index is missing, rebuild it without repacking artifacts:
-
-```bash
-ctx-wiki-graphify \
-  --wiki-dir ~/.claude/skill-wiki \
-  --incremental \
-  --graph-only \
-  --semantic-vector-index numpy-flat
-```
-
-Then drain pending entity-upsert work with `ctx-wiki-worker --wiki
-~/.claude/skill-wiki`. This is the current repair path for "build index" and
-"attach pending" without adding another command surface.
-
-Before publishing graph artifacts, run the full rebuild/export path:
-
-```bash
-ctx-wiki-graphify          # rebuild entity graph + communities
-```
-
-The pre-commit hook (`.githooks/pre-commit`) does **not** rebuild or
-repack graph artifacts from `~/.claude/skill-wiki/`; that local wiki can
-contain private entities. It refreshes cheap README stats when relevant
-checked-in files are staged and warns when entity sources changed. Run
-`ctx-wiki-graphify`, validate, repack, and stage the artifacts explicitly
-for skill, agent, MCP server, or harness releases.
-
-Graphify exports stage and validate each generated artifact before atomic
-promotion. `graph.json`, `graph-delta.json`, `communities.json`,
-`graph-report.md`, and `graph-export-manifest.json` each get a sibling
-`*.promotion.json` file with candidate, current, and `last_good` hashes plus
-rollback metadata. The manifest is promoted last, so a crash between artifact
-promotion and manifest promotion is detected as an incomplete export and the
-next run rebuilds instead of trusting mixed graph files.
-
-## Edge-count history
-
-| Version | Edges | Note |
-|---|---|---|
-| v0.5.x | 642K (stale) / 861 (live) | Bundle had stale 642K; live rebuild silently produced 861 because `DENSE_TAG_THRESHOLD=20` dropped every popular tag. |
-| v0.6.0 | 454,719 | Threshold raised to 500, multi-line YAML lists parsed, slug-token pseudo-tags added. |
-| v0.7.x | 847,207 | Pulsemcp ingest added 10,786 MCP server nodes; sentence-embedding semantic edges added. |
-| 2026-04-27 graph rebuild pass | **963,068** | +21 mattpocock skills, +156 designdotmd designs (+106,702 edges); patch-path bug fixed (graphify now forces full rebuild when prior graph has 0 semantic edges but current run computed semantic pairs); community detection switched from CNM to Louvain. |
-| 2026-04-29 bulk skill import pass | **1,030,831** | +90,846 first-class `skill` nodes, +90,846 skill pages, and +67,519 sparse duplicate/tag metadata edges to the core graph. Full-body semantic edges are intentionally deferred to the hydration pass. |
-| 2026-04-29 text-to-cad harness pass | **1,031,011** | +1 first-class `harness` node, +1 harness page, and +224 explainable harness edges, including 44 skill edges. |
-| 2026-04-29 harness inventory pass | **1,033,253** | +12 first-class `harness` nodes/pages for LangGraph, CrewAI, AutoGen, Google ADK, Semantic Kernel, Mastra, Pydantic AI, Haystack, OpenAI Agents SDK, LiteLLM, Langfuse, and AgentOps; harness incident edges now total 2,700. |
-| 2026-04-30 skill semantic hydration pass | **2,881,027** | +full-body semantic edges for hydrated skill records; semantic top-K became the dominant large-scale signal. |
-| 2026-05-01 micro-skill pass | **2,960,189** | Enforced the <=180-line loader threshold across 89,461 hydrated `SKILL.md` files, converted 28,611 long bodies into gated micro-skill orchestrators, used full originals for semantic graphing, excluded `.original` backups from the shipped tarball, bounded generated stage/reference files to 40 lines, and rebuilt the graph. |
-| 2026-05-02 GitNexus MCP pass | **2,960,215** | Added GitNexus as an MCP server entity with 26 cross-type edges to related skill pages and architecture/refactoring agents; semantic edge count unchanged. |
-| 2026-05-04 v0.7.3 artifact refresh | **2,960,215** | Hydrated one recoverable command-injection-testing body, raising hydrated `SKILL.md` files to 89,463; generated micro-skill markdown now defangs high-risk command-injection payloads before packaging. Graph topology unchanged. |
-| 2026-05-04 body-backed skill prune | **2,900,834** | Removed 1,383 skill records that had no packaged `SKILL.md` body and no parseable prose body. Remaining skill records, graph nodes, entity pages, and converted `SKILL.md` bodies are all **89,463**. |
-| 2026-05-05 artifact hygiene refresh | **2,900,834** | Repacked `graph/wiki-graph.tar.gz` to remove transient `.lock` files from the shipped LLM-wiki. Topology unchanged. |
-| 2026-05-10 v1.0.0 release prep | **2,900,834** | Refreshed shipped HTML previews from the current export, validated their export IDs in CI, and removed stale PNG previews. Topology unchanged. |
-| 2026-05-12 book-to-skill + queue hygiene | **2,900,910** | Added `book-to-skill` as a skill entity (+1 node, +76 edges), restored a missing converted skill body, and repacked `graph/wiki-graph.tar.gz` to omit `.ctx/` queue state. Tar members: **598,154**. |
-| 2026-05-13 overlay pass | **2,911,220** | Added AGENTS.md, lat.md, OptiLLM, Matt Pocock refresh deltas, and Julius caveman entities through the safe overlay path (+21 nodes, +10,310 edges) while preserving the saturated skill topology. Tar members at that pass: **598,192**. |
-| 2026-05-14 Matt Pocock upstream refresh | **2,911,126** | Pinned `mattpocock/skills` to `e74f0061bb67222181640effa98c675bdb2fdaa7`, removed three stale legacy alias skill pages/nodes (`mattpocock-domain-model`, `mattpocock-github-triage`, `mattpocock-triage-issue`), refreshed `mattpocock-grill-with-docs`, and pruned 94 incident edges plus stale wiki references. Current tar members: **598,189**. |
-| 2026-05-14 Mirage + CodeGraph first-class wiki pass | **2,911,162** | Added Mirage as a shipped harness wiki/runtime page, added the CodeGraph MCP markdown page and `codegraph-agentic-codebase-analysis` skill page/body to the full LLM-wiki, added the compact dashboard neighborhood index, regenerated graph preview HTML from the current export, and refreshed exact validation counts. Current tar members: **598,193**. |
-| 2026-05-18 repo refresh | **2,911,575** | Refreshed `addyosmani/agent-skills` and `bytedance/deer-flow` from current upstream SKILL.md bodies, added Addy Osmani's `doubt-driven-development` and `interview-me` skills, replaced DeerFlow's stale `vercel-deploy` entry with `vercel-deploy-claimable`, retained `Imbad0202/academic-research-skills` as existing non-commercial upstream content, and added DeerFlow as a first-class harness page/runtime page. Current tar members: **598,596**. |
-| 2026-05-26 upstream repo ingest | **2,913,930** | Added `browsing-skills/browsing-skills` as a root browser skill plus 11 site-specific browser skills, refreshed `awesome-skills/code-review-skill` into `code-review-excellence` with its reference pack, refreshed `book-to-skill` through the micro-skill gate, added Presenton as both a harness and MCP server, refreshed DeerFlow's harness page, imported 190 harness catalog pages from `Picrew/awesome-agent-harness`, regenerated dashboard sqlite and preview HTML from the current export, and refreshed exact validation counts. Current tar members: **598,951**. |
-| 2026-05-27 Nango catalog/tooling overlay | **2,913,960** | Added `nango-integration-catalog` as a skill page/body, added Nango hosted action-function MCP and Nango docs MCP as first-class MCP pages, connected them to API/auth/tool-calling skills, common SaaS MCPs, and agent harnesses, regenerated dashboard sqlite and preview HTML, and refreshed exact validation counts. Current tar members: **598,955**. |
-| 2026-05-27 Nango semantic rescore | **2,913,960** | Replaced Nango's placeholder manual edge weights with measured MiniLM semantic cosines plus graphify-compatible tag/direct/type components. Topology unchanged; semantic edge count increased by 30 to **1,683,193**. |
-
-The full audit history lives in `CHANGELOG.md`. The current build is
-fully reproducible from the wiki content.
-
-## Pre-ship gates
-
-Two advisory gates run before the tarball is repackaged. Both produce
-review reports and never auto-modify the inventory.
-
-- **`ctx-dedup-check`** — flags entity pairs (skill ↔ skill, skill ↔
-  agent, skill ↔ MCP, agent ↔ agent, agent ↔ MCP, MCP ↔ MCP) at or
-  above 0.85 cosine similarity. Incremental: keeps a `dedup-state.json`
-  next to the embedding cache, so follow-up runs only re-check pairs
-  involving entities whose content changed. Allowlist support via
-  `.dedup-allowlist.txt`. The current snapshot has 15,976 findings,
-  most of which are within-MCP near-duplicates (multiple wrappers
-  around the same upstream service).
-- **`ctx-tag-backfill`** — finds skills/agents with empty `tags:`
-  frontmatter and proposes a backfill drawn from slug tokens, body
-  keywords, and the existing tag vocabulary. Report-only by default;
-  pass `--apply` to write. Backfills are additive only.
+# Knowledge graph
+
+A pre-built weighted graph of skills, agents, MCP servers, and
+harnesses in the ctx ecosystem, shipped as `graph/wiki-graph.tar.gz`.
+The on-disk JSON and `resolve_graph` Python API are harness-aware, including
+plain-slug graph walks from `harness:<slug>` nodes. `ctx-monitor`
+exposes skill/agent/MCP/harness wiki and graph views. Harness installation,
+update, and uninstall are handled by `ctx-harness-install`; dashboard
+load/unload POSTs deliberately reject harnesses and return the dry-run CLI
+command to use instead. Quality scoring is exposed for sidecar-backed skills,
+agents, and MCP servers.
+
+## What's in it
+
+Authoritative numbers from the shipped tarball. The curated-core snapshot
+is **13,463 nodes** (1,999 curated skills + 467 agents + 10,790 MCP servers + 207 harnesses). Harness pages under `entities/harnesses/` are ingested into
+local rebuilds and the separate harness recommendation path. The
+tarball also carries **91,464 skill pages**; **89,465**
+skill bodies are hydrated as installable `SKILL.md` files under
+`converted/`; the **28,612** entries over the configured line
+limit were converted to gated micro-skill orchestrators. Full original bodies
+are used during graph rebuilds for semantic similarity, but
+`SKILL.md.original` backups, transient `.lock` files, and `.ctx/` queue state
+are omitted from the shipped tarball.
+
+| | Count |
+|---|---:|
+| Total nodes | **102,928** |
+| Curated core nodes | **13,463** (1,999 skills + 467 agents + 10,790 MCP servers + 207 harnesses) |
+| Body-backed skill nodes | **89,465** hydrated installable skill entries |
+| Total edges | **2,913,960** |
+| Hydrated skill incident edges | **2,605,721** |
+| Hydrated skill semantic incident edges | **1,500,648** |
+| Communities | **52** (Louvain) |
+| Edge sources (overlap-deduped) | semantic 1,683,193 - tag 897,784 - token 433,245 |
+| Cross-type edges (skill <-> agent) | ~66,799 |
+| Cross-type edges (skill <-> MCP) | ~41,521 |
+| Cross-type edges (agent <-> MCP) | ~229 |
+| Harness edges | **6,576** |
+| Shipped skill index | **89,465** observed body-backed skill entries |
+
+## Install
+
+Use `ctx-init --graph` to install the fast runtime graph. Source checkouts use
+`graph/wiki-graph-runtime.tar.gz`; pip installs download the matching GitHub
+release asset for the installed package version. This installs
+`graphify-out/*`, the skill index used by recommendations, and
+the harness pages used by `ctx-harness-install`:
+
+```bash
+ctx-init --graph
+```
+
+To expand every shipped skill/agent/MCP entity page, harness page,
+skill page, concept page, converted micro-skill pipeline,
+and Obsidian vault metadata, request the full wiki artifact explicitly:
+
+```bash
+ctx-init --graph --graph-install-mode full
+```
+
+Manual extraction is still supported for offline/source installs. Extract the
+full tarball into your `~/.claude/skill-wiki/` when you want local markdown
+wiki browsing:
+
+```bash
+mkdir -p ~/.claude/skill-wiki
+tar xzf graph/wiki-graph.tar.gz -C ~/.claude/skill-wiki/
+```
+
+On Windows PowerShell, create the target and use the built-in `tar.exe`
+without `--force-local`:
+
+```powershell
+New-Item -ItemType Directory -Force "$env:USERPROFILE\.claude\skill-wiki" | Out-Null
+tar -xzf graph\wiki-graph.tar.gz -C "$env:USERPROFILE\.claude\skill-wiki"
+```
+
+The extracted tree also opens directly as an Obsidian vault — the
+`.obsidian/` config ships inside the tarball — so you can use
+Obsidian's native graph view if you prefer it to the web dashboard.
+
+## How edges are built
+
+Edges are built and explained by the `ctx-wiki-graphify` console script
+(`ctx.core.wiki.wiki_graphify`). A pair must first have at least one base
+signal:
+
+1. **Semantic cosine** — when the embedding backend is available, entity
+   text is embedded and semantic neighbors above the configured build floor
+   contribute weighted edges.
+2. **Explicit frontmatter tags** — each entity page's YAML `tags:`
+   list contributes edges between every pair of entities that share
+   a tag. Popular tags capped at 500 nodes to avoid noise-floor
+   "everything connects to everything" mega-buckets like `typescript`
+   or `frontend`.
+3. **Slug-token pseudo-tags** — each hyphenated slug contributes its
+   tokens as implicit tags. `fastapi-pro` contributes `fastapi`;
+   `python-patterns` contributes `python` and `patterns`. A stop-word
+   filter drops generic tokens like `skill`, `agent`, `pro`, `expert`,
+   `core` so they don't over-connect the graph.
+4. **Source overlap** — pages with the same high-specificity source URL,
+   repository URL, homepage, detail URL, or package URL can connect even
+   when their tags differ. Dense source buckets are skipped.
+5. **Direct wikilinks** — explicit entity links such as
+   `[[entities/agents/code-reviewer]]` create a direct graph edge.
+
+Edge `weight` is the final blended strength. Semantic, tag, and token
+weights form the base blend from `config.json`; source overlap and direct
+links add configured boosts. Existing edges can also receive explainable
+ranking boosts from Adamic-Adar shared-neighbor structure, type affinity,
+usage telemetry, and quality scores. Those boost-only signals do not create
+edges by themselves. The shipped default `graph.min_edge_weight` is `0.03`;
+calibration against the 2026-05 shipped graph showed this is the highest
+floor with zero edge loss, while `0.05` would remove roughly 29.7% of edges.
+
+Edge metadata keeps the ingredients explainable: `semantic_sim`,
+`shared_tags`, `shared_tokens`, `shared_sources`, `direct_link`,
+`adamic_adar`, `type_affinity`, `usage_score`, `quality_score`,
+`edge_reasons`, and `score_components`. Hydrated skill records use their
+full source bodies during graph rebuilds, so long converted entries keep
+full-body similarity even though the shipped installable `SKILL.md` files are
+short gated loaders. The raw `SKILL.md.original` backups are build inputs, not
+tarball members.
+
+## Communities
+
+After edges are built, `wiki_graphify` runs NetworkX's Louvain
+community detection (`resolution=1.2`, `seed=42` for determinism).
+The result is **52 communities** ranging from single-member isolated
+specialists to several thousand members in broad clusters like
+`Community + Official + AI`. Each community also gets an auto-generated
+`concepts/<community>.md` wiki page summarizing its members and top
+shared tags.
+
+The legacy CNM ("greedy modularity") algorithm is still available
+behind `CTX_GRAPH_COMMUNITY=cnm` — it's deterministic but O(n²) on
+dense graphs and hangs on the live 13K-node dataset (~50min run was
+killed on 2026-04-27 inside the priority-queue siftup). Louvain is
+the default because it finishes in seconds and produces equivalent
+quality clusters for the recommendation use case.
+
+## Querying the graph
+
+### Via the dashboard
+
+```bash
+ctx-monitor serve              # http://127.0.0.1:8765
+```
+
+Then open `/graph?slug=<entity-slug>&type=<entity-type>` for a
+cytoscape neighborhood view, or
+`/api/graph/<slug>.json?type=<entity-type>&hops=1&limit=40` for the
+dashboard-shaped JSON. The `type` query is optional for unique slugs and
+recommended for duplicate slugs such as `langgraph`. See the
+[dashboard reference](dashboard.md) for the full route catalogue.
+
+### Via Python
+
+```python
+import json
+from pathlib import Path
+from networkx.readwrite import node_link_graph
+
+raw = json.loads(
+    Path("~/.claude/skill-wiki/graphify-out/graph.json").expanduser().read_text()
+)
+edges_key = "links" if "links" in raw else "edges"
+G = node_link_graph(raw, edges=edges_key)
+
+# 102,928 nodes, 2,913,960 edges
+print(G.number_of_nodes(), G.number_of_edges())
+
+# Find entities related to 'fastapi-pro' by edge weight
+seed = "skill:fastapi-pro"
+neighbors = sorted(
+    G.neighbors(seed),
+    key=lambda n: G[seed][n]["weight"],
+    reverse=True,
+)[:10]
+for n in neighbors:
+    shared = G[seed][n].get("shared_tags", [])
+    print(f"  w={G[seed][n]['weight']:>2}  {G.nodes[n]['label']:<40}  {shared[:3]}")
+```
+
+The node-link JSON schema's edges key is auto-detected (legacy
+NetworkX 2.x used `"links"`; current versions default to `"edges"`).
+The helper `resolve_graph.load_graph()` does this for you.
+
+### Via recommendation paths
+
+The graph backs two recommendation paths:
+
+- Execution recommendation surfaces (`ctx.recommend_bundle`, MCP
+  `ctx__recommend_bundle`, generic harness tools, Claude Code hook
+  suggestions, and repo-scan advisory output) share
+  `ctx.core.resolve.recommendations.recommend_by_tags` for skills,
+  agents, and MCP servers. That engine ranks candidates by
+  slug-token matches, tag overlap, graph degree, and semantic-cache
+  signals when available. Imported skill results are normal `skill` nodes with
+  detail URLs, install commands, duplicate
+  hints, gated micro-skill loaders when over the line threshold, and
+  quality/security metadata. If an older
+  extracted wiki has the skill index JSON but no graph nodes for
+  those records, the same recommender falls back to the index file.
+- Harness recommendations are a separate path for custom/API/local
+  model onboarding (`ctx-init --model-mode custom ...`) and
+  `ctx-harness-install`. They use the same graph filtered to
+  `harness` nodes and the higher harness match floor from `config.json`.
+- Repository scans still start from stack detections, then turn that profile
+  into the same tag/query bundle used by the execution recommender. If a
+  shipped graph is unavailable, scan output falls back to the legacy installed
+  skill resolver so a plain profile scan remains useful. Harnesses are
+  intentionally not emitted from repo scans or Claude Code hook bundles.
+
+This split is intentional: execution surfaces need identical ranking and a
+small top-K, while harness choice changes the model runtime itself and belongs
+in an explicit onboarding/install flow.
+
+### LLM-wiki design references
+
+ctx follows Karpathy's LLM-wiki pattern. We also reviewed
+[`nashsu/llm_wiki`](https://github.com/nashsu/llm_wiki) as a design reference
+for source traceability, persistent ingest queues, graph insights, and
+budgeted token/vector/graph retrieval. That repository is GPLv3, while ctx is
+MIT, so ctx can use those ideas as product inspiration but must not copy or
+vendor its code or assets.
+
+## Rebuilding
+
+After you add a skill, agent, MCP server, or harness entity page:
+
+```bash
+ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1
+```
+
+The `entity-upsert` worker path validates the queued page hash, updates the
+wiki index, and, when a persisted semantic vector index exists, runs a
+best-effort ANN attach into `graphify-out/entity-overlays.jsonl`. That overlay
+lets the runtime resolver connect a new or updated entity to existing graph
+neighbors without recomputing global all-pairs similarity. The worker still
+queues the normal incremental `graph-export` job, and the entity markdown page
+remains the source of truth.
+
+For manual review or debugging:
+
+```bash
+ctx-incremental-attach calibrate \
+  --graph ~/.claude/skill-wiki/graphify-out/graph.json
+
+ctx-incremental-attach attach \
+  --index-dir ~/.claude/skill-wiki/.embedding-cache/graph/vector-index \
+  --overlay ~/.claude/skill-wiki/graphify-out/entity-overlays.jsonl \
+  --node-id skill:fastapi-review \
+  --type skill \
+  --label fastapi-review \
+  --text-file ~/.claude/skill-wiki/entities/skills/fastapi-review.md \
+  --dry-run
+```
+
+Shadow-gate a persisted index before trusting a new ANN backend, changed
+thresholds, or a large attach workflow:
+
+```bash
+ctx-incremental-shadow \
+  --index-dir ~/.claude/skill-wiki/.embedding-cache/graph/vector-index \
+  --graph ~/.claude/skill-wiki/graphify-out/graph.json \
+  --sample-size 100 \
+  --min-overlap 0.85
+```
+
+The shadow command pretends sampled existing nodes are new, compares the
+incremental attach result to batch graph semantic neighbors, and reports
+precision, recall, top-5/top-10/top-20 agreement, score deltas, and bad
+examples. A failing gate means either tune thresholds or use a full graph
+rebuild before shipping.
+
+If the vector index is missing, rebuild it without repacking artifacts:
+
+```bash
+ctx-wiki-graphify \
+  --wiki-dir ~/.claude/skill-wiki \
+  --incremental \
+  --graph-only \
+  --semantic-vector-index numpy-flat
+```
+
+Then drain pending entity-upsert work with `ctx-wiki-worker --wiki
+~/.claude/skill-wiki`. This is the current repair path for "build index" and
+"attach pending" without adding another command surface.
+
+Before publishing graph artifacts, run the full rebuild/export path:
+
+```bash
+ctx-wiki-graphify          # rebuild entity graph + communities
+```
+
+The pre-commit hook (`.githooks/pre-commit`) does **not** rebuild or
+repack graph artifacts from `~/.claude/skill-wiki/`; that local wiki can
+contain private entities. It refreshes cheap README stats when relevant
+checked-in files are staged and warns when entity sources changed. Run
+`ctx-wiki-graphify`, validate, repack, and stage the artifacts explicitly
+for skill, agent, MCP server, or harness releases.
+
+Graphify exports stage and validate each generated artifact before atomic
+promotion. `graph.json`, `graph-delta.json`, `communities.json`,
+`graph-report.md`, and `graph-export-manifest.json` each get a sibling
+`*.promotion.json` file with candidate, current, and `last_good` hashes plus
+rollback metadata. The manifest is promoted last, so a crash between artifact
+promotion and manifest promotion is detected as an incomplete export and the
+next run rebuilds instead of trusting mixed graph files.
+
+## Edge-count history
+
+| Version | Edges | Note |
+|---|---|---|
+| v0.5.x | 642K (stale) / 861 (live) | Bundle had stale 642K; live rebuild silently produced 861 because `DENSE_TAG_THRESHOLD=20` dropped every popular tag. |
+| v0.6.0 | 454,719 | Threshold raised to 500, multi-line YAML lists parsed, slug-token pseudo-tags added. |
+| v0.7.x | 847,207 | Pulsemcp ingest added 10,786 MCP server nodes; sentence-embedding semantic edges added. |
+| 2026-04-27 graph rebuild pass | **963,068** | +21 mattpocock skills, +156 designdotmd designs (+106,702 edges); patch-path bug fixed (graphify now forces full rebuild when prior graph has 0 semantic edges but current run computed semantic pairs); community detection switched from CNM to Louvain. |
+| 2026-04-29 bulk skill import pass | **1,030,831** | +90,846 first-class `skill` nodes, +90,846 skill pages, and +67,519 sparse duplicate/tag metadata edges to the core graph. Full-body semantic edges are intentionally deferred to the hydration pass. |
+| 2026-04-29 text-to-cad harness pass | **1,031,011** | +1 first-class `harness` node, +1 harness page, and +224 explainable harness edges, including 44 skill edges. |
+| 2026-04-29 harness inventory pass | **1,033,253** | +12 first-class `harness` nodes/pages for LangGraph, CrewAI, AutoGen, Google ADK, Semantic Kernel, Mastra, Pydantic AI, Haystack, OpenAI Agents SDK, LiteLLM, Langfuse, and AgentOps; harness incident edges now total 2,700. |
+| 2026-04-30 skill semantic hydration pass | **2,881,027** | +full-body semantic edges for hydrated skill records; semantic top-K became the dominant large-scale signal. |
+| 2026-05-01 micro-skill pass | **2,960,189** | Enforced the <=180-line loader threshold across 89,461 hydrated `SKILL.md` files, converted 28,611 long bodies into gated micro-skill orchestrators, used full originals for semantic graphing, excluded `.original` backups from the shipped tarball, bounded generated stage/reference files to 40 lines, and rebuilt the graph. |
+| 2026-05-02 GitNexus MCP pass | **2,960,215** | Added GitNexus as an MCP server entity with 26 cross-type edges to related skill pages and architecture/refactoring agents; semantic edge count unchanged. |
+| 2026-05-04 v0.7.3 artifact refresh | **2,960,215** | Hydrated one recoverable command-injection-testing body, raising hydrated `SKILL.md` files to 89,463; generated micro-skill markdown now defangs high-risk command-injection payloads before packaging. Graph topology unchanged. |
+| 2026-05-04 body-backed skill prune | **2,900,834** | Removed 1,383 skill records that had no packaged `SKILL.md` body and no parseable prose body. Remaining skill records, graph nodes, entity pages, and converted `SKILL.md` bodies are all **89,463**. |
+| 2026-05-05 artifact hygiene refresh | **2,900,834** | Repacked `graph/wiki-graph.tar.gz` to remove transient `.lock` files from the shipped LLM-wiki. Topology unchanged. |
+| 2026-05-10 v1.0.0 release prep | **2,900,834** | Refreshed shipped HTML previews from the current export, validated their export IDs in CI, and removed stale PNG previews. Topology unchanged. |
+| 2026-05-12 book-to-skill + queue hygiene | **2,900,910** | Added `book-to-skill` as a skill entity (+1 node, +76 edges), restored a missing converted skill body, and repacked `graph/wiki-graph.tar.gz` to omit `.ctx/` queue state. Tar members: **598,154**. |
+| 2026-05-13 overlay pass | **2,911,220** | Added AGENTS.md, lat.md, OptiLLM, Matt Pocock refresh deltas, and Julius caveman entities through the safe overlay path (+21 nodes, +10,310 edges) while preserving the saturated skill topology. Tar members at that pass: **598,192**. |
+| 2026-05-14 Matt Pocock upstream refresh | **2,911,126** | Pinned `mattpocock/skills` to `e74f0061bb67222181640effa98c675bdb2fdaa7`, removed three stale legacy alias skill pages/nodes (`mattpocock-domain-model`, `mattpocock-github-triage`, `mattpocock-triage-issue`), refreshed `mattpocock-grill-with-docs`, and pruned 94 incident edges plus stale wiki references. Current tar members: **598,189**. |
+| 2026-05-14 Mirage + CodeGraph first-class wiki pass | **2,911,162** | Added Mirage as a shipped harness wiki/runtime page, added the CodeGraph MCP markdown page and `codegraph-agentic-codebase-analysis` skill page/body to the full LLM-wiki, added the compact dashboard neighborhood index, regenerated graph preview HTML from the current export, and refreshed exact validation counts. Current tar members: **598,193**. |
+| 2026-05-18 repo refresh | **2,911,575** | Refreshed `addyosmani/agent-skills` and `bytedance/deer-flow` from current upstream SKILL.md bodies, added Addy Osmani's `doubt-driven-development` and `interview-me` skills, replaced DeerFlow's stale `vercel-deploy` entry with `vercel-deploy-claimable`, retained `Imbad0202/academic-research-skills` as existing non-commercial upstream content, and added DeerFlow as a first-class harness page/runtime page. Current tar members: **598,596**. |
+| 2026-05-26 upstream repo ingest | **2,913,930** | Added `browsing-skills/browsing-skills` as a root browser skill plus 11 site-specific browser skills, refreshed `awesome-skills/code-review-skill` into `code-review-excellence` with its reference pack, refreshed `book-to-skill` through the micro-skill gate, added Presenton as both a harness and MCP server, refreshed DeerFlow's harness page, imported 190 harness catalog pages from `Picrew/awesome-agent-harness`, regenerated dashboard sqlite and preview HTML from the current export, and refreshed exact validation counts. Current tar members: **598,951**. |
+| 2026-05-27 Nango catalog/tooling overlay | **2,913,960** | Added `nango-integration-catalog` as a skill page/body, added Nango hosted action-function MCP and Nango docs MCP as first-class MCP pages, connected them to API/auth/tool-calling skills, common SaaS MCPs, and agent harnesses, regenerated dashboard sqlite and preview HTML, and refreshed exact validation counts. Current tar members: **598,955**. |
+| 2026-05-27 Nango semantic rescore | **2,913,960** | Replaced Nango's placeholder manual edge weights with measured MiniLM semantic cosines plus graphify-compatible tag/direct/type components. Topology unchanged; semantic edge count increased by 30 to **1,683,193**. |
+
+The full audit history lives in `CHANGELOG.md`. The current build is
+fully reproducible from the wiki content.
+
+## Pre-ship gates
+
+Two advisory gates run before the tarball is repackaged. Both produce
+review reports and never auto-modify the inventory.
+
+- **`ctx-dedup-check`** — flags entity pairs (skill ↔ skill, skill ↔
+  agent, skill ↔ MCP, agent ↔ agent, agent ↔ MCP, MCP ↔ MCP) at or
+  above 0.85 cosine similarity. Incremental: keeps a `dedup-state.json`
+  next to the embedding cache, so follow-up runs only re-check pairs
+  involving entities whose content changed. Allowlist support via
+  `.dedup-allowlist.txt`. The current snapshot has 15,976 findings,
+  most of which are within-MCP near-duplicates (multiple wrappers
+  around the same upstream service).
+- **`ctx-tag-backfill`** — finds skills/agents with empty `tags:`
+  frontmatter and proposes a backfill drawn from slug tokens, body
+  keywords, and the existing tag vocabulary. Report-only by default;
+  pass `--apply` to write. Backfills are additive only.

docs/marketplace-registry.md

@@ -1,129 +1,129 @@
-# Entity Source Registry
-
-ctx keeps entity sources separate from install state. A source entry can be
-searched and recommended without being installed, and duplicate/update paths must
-show what would change before replacing an existing entity.
-
-## Registered Sources
-
-### Shipped Graph And LLM-Wiki
-
-```yaml
-name: ctx-shipped-graph
-type: compressed-runtime
-paths:
-  graph: graph/wiki-graph-runtime.tar.gz
-  skills: graph/<skill-index>.json.gz
-refresh: release-time
-priority: 1
-```
-
-The shipped runtime is the first source for recommendations. It contains the
-first-class skills, agents, MCP servers, harnesses, graph edges, quality scores,
-usage fields, and wiki pages that ctx can use offline.
-
-### User Local Assets
-
-```yaml
-name: user-local
-type: filesystem
-paths:
-  skills: ~/.claude/skills
-  agents: ~/.claude/agents
-  mcp: project/user MCP config files
-  harnesses: ~/.ctx/harnesses
-refresh: always current
-priority: 2
-```
-
-Local assets override shipped suggestions when names collide, but updates
-still require an explicit review if replacement content is proposed.
-
-### Skill Index
-
-```yaml
-name: shipped-skills
-type: shipped-index
-shipped_entries: 89465
-local_index: graph/<skill-index>.json.gz
-hydrated_wiki: external-catalogs/<source>/catalog.json
-refresh: on-demand
-priority: 3
-```
-
-Skill index entries are stored as first-class skill entities in the graph/wiki.
-Hydrated `SKILL.md` bodies pass through the micro-skill gate before they are
-packed into the shipped runtime.
-
-### GitHub Entity Repositories
-
-```yaml
-name: github-entity-repos
-type: git
-entrypoints:
-  skills: ctx-skill-add
-  agents: ctx-agent-add
-  mcp: ctx-mcp-add
-  harnesses: ctx-harness-add
-refresh: on-demand
-priority: 4
-```
-
-GitHub stars, forks, releases, and update timestamps can be stored as source
-metadata for a candidate entity. The ctx repository's own star count is not
-hard-coded in docs because it changes continuously; read it from GitHub when
-needed.
-
-### MCP And Harness Sources
-
-```yaml
-name: mcp-and-harness-sources
-type: curated-source
-entrypoints:
-  mcp: ctx-mcp-fetch, ctx-mcp-add
-  harnesses: ctx-harness-add, ctx-harness-install
-refresh: on-demand
-priority: 5
-```
-
-MCP servers and harnesses are recorded as entities with install guidance,
-permission notes, compatibility tags, and quality/security review status.
-
-## Query Protocol
-
-When a user asks for help or the scanner detects a stack/task gap:
-
-1. Search the local graph/wiki first.
-2. Search the shipped skill index and, when needed, the `find-skills`
-   helper for fresher remote results.
-3. Search local user assets and configured entity repositories.
-4. Deduplicate by slug, source URL, canonical name, tags, and semantic overlap.
-5. Score candidates with the shared recommendation engine.
-6. Present at most the configured cap with reasons, quality/usage notes, and
-   install/update commands.
-7. If an existing entity would be replaced, show the update review: benefits,
-   drawbacks, changed files, security posture, and rollback path.
-
-## Update Rules
-
-Entity updates are intentionally explicit:
-
-- `ctx-skill-add`, `ctx-agent-add`, `ctx-mcp-add`, and `ctx-harness-add` create
-  new entities when no duplicate exists.
-- If a duplicate exists, the command emits an update review and refuses to
-  replace content unless the user passes the update flag.
-- New or updated skills go through the micro-skill line-count gate from config.
-- Security/cyber checks run before entity content is promoted.
-- Graph/wiki artifacts are rebuilt, validated, packed, and atomically promoted.
-
-## Security Notes
-
-- Never auto-install without user confirmation.
-- Always show the source URL, entity type, install command, and permissions.
-- Reject missing `SKILL.md` bodies for skill imports unless the source is only a
-  catalog pointer.
-- Never execute repository scripts during cataloging without explicit user
-  approval.
-- Warn on network, filesystem, shell, credential, or system-level permissions.
-- Preserve last-good graph/wiki artifacts so a failed refresh cannot ship a
-  corrupt runtime.
+# Entity Source Registry
+
+ctx keeps entity sources separate from install state. A source entry can be
+searched and recommended without being installed, and duplicate/update paths must
+show what would change before replacing an existing entity.
+
+## Registered Sources
+
+### Shipped Graph And LLM-Wiki
+
+```yaml
+name: ctx-shipped-graph
+type: compressed-runtime
+paths:
+  graph: graph/wiki-graph-runtime.tar.gz
+  skills: graph/<skill-index>.json.gz
+refresh: release-time
+priority: 1
+```
+
+The shipped runtime is the first source for recommendations. It contains the
+first-class skills, agents, MCP servers, harnesses, graph edges, quality scores,
+usage fields, and wiki pages that ctx can use offline.
+
+### User Local Assets
+
+```yaml
+name: user-local
+type: filesystem
+paths:
+  skills: ~/.claude/skills
+  agents: ~/.claude/agents
+  mcp: project/user MCP config files
+  harnesses: ~/.ctx/harnesses
+refresh: always current
+priority: 2
+```
+
+Local assets override shipped suggestions when names collide, but updates
+still require an explicit review if replacement content is proposed.
+
+### Skill Index
+
+```yaml
+name: shipped-skills
+type: shipped-index
+shipped_entries: 89465
+local_index: graph/<skill-index>.json.gz
+hydrated_wiki: external-catalogs/<source>/catalog.json
+refresh: on-demand
+priority: 3
+```
+
+Skill index entries are stored as first-class skill entities in the graph/wiki.
+Hydrated `SKILL.md` bodies pass through the micro-skill gate before they are
+packed into the shipped runtime.
+
+### GitHub Entity Repositories
+
+```yaml
+name: github-entity-repos
+type: git
+entrypoints:
+  skills: ctx-skill-add
+  agents: ctx-agent-add
+  mcp: ctx-mcp-add
+  harnesses: ctx-harness-add
+refresh: on-demand
+priority: 4
+```
+
+GitHub stars, forks, releases, and update timestamps can be stored as source
+metadata for a candidate entity. The ctx repository's own star count is not
+hard-coded in docs because it changes continuously; read it from GitHub when
+needed.
+
+### MCP And Harness Sources
+
+```yaml
+name: mcp-and-harness-sources
+type: curated-source
+entrypoints:
+  mcp: ctx-mcp-fetch, ctx-mcp-add
+  harnesses: ctx-harness-add, ctx-harness-install
+refresh: on-demand
+priority: 5
+```
+
+MCP servers and harnesses are recorded as entities with install guidance,
+permission notes, compatibility tags, and quality/security review status.
+
+## Query Protocol
+
+When a user asks for help or the scanner detects a stack/task gap:
+
+1. Search the local graph/wiki first.
+2. Search the shipped skill index and, when needed, the `find-skills`
+   helper for fresher remote results.
+3. Search local user assets and configured entity repositories.
+4. Deduplicate by slug, source URL, canonical name, tags, and semantic overlap.
+5. Score candidates with the shared recommendation engine.
+6. Present at most the configured cap with reasons, quality/usage notes, and
+   install/update commands.
+7. If an existing entity would be replaced, show the update review: benefits,
+   drawbacks, changed files, security posture, and rollback path.
+
+## Update Rules
+
+Entity updates are intentionally explicit:
+
+- `ctx-skill-add`, `ctx-agent-add`, `ctx-mcp-add`, and `ctx-harness-add` create
+  new entities when no duplicate exists.
+- If a duplicate exists, the command emits an update review and refuses to
+  replace content unless the user passes the update flag.
+- New or updated skills go through the micro-skill line-count gate from config.
+- Security/cyber checks run before entity content is promoted.
+- Graph/wiki artifacts are rebuilt, validated, packed, and atomically promoted.
+
+## Security Notes
+
+- Never auto-install without user confirmation.
+- Always show the source URL, entity type, install command, and permissions.
+- Reject missing `SKILL.md` bodies for skill imports unless the source is only a
+  catalog pointer.
+- Never execute repository scripts during cataloging without explicit user
+  approval.
+- Warn on network, filesystem, shell, credential, or system-level permissions.
+- Preserve last-good graph/wiki artifacts so a failed refresh cannot ship a
+  corrupt runtime.

docs/memory-anchor.md

@@ -1,102 +1,102 @@
-# Memory anchoring
-
-[`src/memory_anchor.py`](https://github.com/stevesolun/ctx/blob/main/src/memory_anchor.py)
-walks the auto-memory store and flags references that no longer
-resolve against the current repository.
-
-## Why it exists
-
-Claude's auto-memory accumulates notes that look like:
-
-> Fixed the `add_skill()` bug in `src/skill_loader.py:42`. See also
-> `docs/intent-interview.md`.
-
-Those backtick references rot as the codebase moves. A renamed file, a
-deleted module, a moved doc — and the memory silently points at
-nothing. `memory_anchor` turns that silent rot into a loud dashboard.
-
-## Where it looks
-
-Memory files live under:
-
-```text
-~/.claude/projects/<slug>/memory/*.md
-```
-
-The module recursively scans that tree. You can override the root with
-`--memory-root` (useful for tests or multi-project setups).
-
-## What counts as a reference
-
-Only tokens inside **backtick code spans** qualify. The heuristic is
-deliberately conservative to keep false positives low:
-
-- known extension (`.py`, `.md`, `.json`, `.yml`, `.ts`, `.rs`, …), or
-- contains a `/` with a dotted final segment.
-
-Tokens with whitespace, `()` suffixes, `http(s)://` prefixes, or leading
-`-` are rejected up front. A trailing `:<digits>` is parsed as a line
-suffix — `:` without digits is preserved (keeps Windows drive letters
-intact).
-
-## How resolution works
-
-For each extracted reference, the module asks whether it resolves:
-
-1. tilde-expand, if `~/…`
-2. if absolute, does the path exist?
-3. does `repo_root / path` exist?
-4. does `repo_root / src / path` exist?
-
-If any candidate hits, the ref is *live*; otherwise *dead*.
-
-## CLI
-
-```bash
-# JSON report for downstream tooling
-python -m memory_anchor scan
-
-# Human dashboard
-python -m memory_anchor dashboard
-
-# CI gate: exit 2 if any dead references remain
-python -m memory_anchor check --strict
-
-# Override repo / memory roots
-python -m memory_anchor check --strict \
-  --repo-root /path/to/repo \
-  --memory-root /path/to/project/memory
-```
-
-When `--repo-root` is omitted, the module walks upward from the current
-directory to the nearest `.git/` ancestor.
-
-## Data model
-
-```python
-@dataclass(frozen=True)
-class AnchorRef:
-    raw: str          # exactly the backtick contents
-    path: str         # path sans trailing :<line>
-    line: int | None
-    exists: bool
-
-@dataclass(frozen=True)
-class MemoryAnchorFile:
-    memory_path: str
-    refs: tuple[AnchorRef, ...]
-    # derived: .live, .dead
-
-@dataclass(frozen=True)
-class AnchorReport:
-    generated_at: float
-    repo_root: str
-    memory_root: str
-    files: tuple[MemoryAnchorFile, ...]
-    # derived: .all_refs, .live_count, .dead_count, .has_dead
-```
-
-## Related
-
-- [Skill health dashboard](skills-health.md) — structural and drift
-  checks for the skill + agent catalog.
+# Memory anchoring
+
+[`src/memory_anchor.py`](https://github.com/stevesolun/ctx/blob/main/src/memory_anchor.py)
+walks the auto-memory store and flags references that no longer
+resolve against the current repository.
+
+## Why it exists
+
+Claude's auto-memory accumulates notes that look like:
+
+> Fixed the `add_skill()` bug in `src/skill_loader.py:42`. See also
+> `docs/intent-interview.md`.
+
+Those backtick references rot as the codebase moves. A renamed file, a
+deleted module, a moved doc — and the memory silently points at
+nothing. `memory_anchor` turns that silent rot into a loud dashboard.
+
+## Where it looks
+
+Memory files live under:
+
+```text
+~/.claude/projects/<slug>/memory/*.md
+```
+
+The module recursively scans that tree. You can override the root with
+`--memory-root` (useful for tests or multi-project setups).
+
+## What counts as a reference
+
+Only tokens inside **backtick code spans** qualify. The heuristic is
+deliberately conservative to keep false positives low:
+
+- known extension (`.py`, `.md`, `.json`, `.yml`, `.ts`, `.rs`, …), or
+- contains a `/` with a dotted final segment.
+
+Tokens with whitespace, `()` suffixes, `http(s)://` prefixes, or leading
+`-` are rejected up front. A trailing `:<digits>` is parsed as a line
+suffix — `:` without digits is preserved (keeps Windows drive letters
+intact).
+
+## How resolution works
+
+For each extracted reference, the module asks whether it resolves:
+
+1. tilde-expand, if `~/…`
+2. if absolute, does the path exist?
+3. does `repo_root / path` exist?
+4. does `repo_root / src / path` exist?
+
+If any candidate hits, the ref is *live*; otherwise *dead*.
+
+## CLI
+
+```bash
+# JSON report for downstream tooling
+python -m memory_anchor scan
+
+# Human dashboard
+python -m memory_anchor dashboard
+
+# CI gate: exit 2 if any dead references remain
+python -m memory_anchor check --strict
+
+# Override repo / memory roots
+python -m memory_anchor check --strict \
+  --repo-root /path/to/repo \
+  --memory-root /path/to/project/memory
+```
+
+When `--repo-root` is omitted, the module walks upward from the current
+directory to the nearest `.git/` ancestor.
+
+## Data model
+
+```python
+@dataclass(frozen=True)
+class AnchorRef:
+    raw: str          # exactly the backtick contents
+    path: str         # path sans trailing :<line>
+    line: int | None
+    exists: bool
+
+@dataclass(frozen=True)
+class MemoryAnchorFile:
+    memory_path: str
+    refs: tuple[AnchorRef, ...]
+    # derived: .live, .dead
+
+@dataclass(frozen=True)
+class AnchorReport:
+    generated_at: float
+    repo_root: str
+    memory_root: str
+    files: tuple[MemoryAnchorFile, ...]
+    # derived: .all_refs, .live_count, .dead_count, .has_dead
+```
+
+## Related
+
+- [Skill health dashboard](skills-health.md) — structural and drift
+  checks for the skill + agent catalog.

docs/skill-lifecycle-and-dashboard.md

@@ -1,171 +1,171 @@
-# Skill lifecycle & KPI dashboard — install & operations
-
-One page on running the lifecycle CLI, the category backfill, and the KPI
-dashboard. Prerequisite: the quality scorer is installed and has written at
-least one sidecar. See
-[skill-quality-install.md](./skill-quality-install.md).
-
-## What it does
-
-After the scorer has labeled everything, these three tools turn the
-labels into action:
-
-| Tool | CLI | Purpose |
-| ---- | --- | ------- |
-| `ctx_lifecycle.py`  | `review`, `demote`, `archive`, `purge`, `review-archived` | Move D/F-grade skills through `active → watch → demote → archive → deleted`. |
-| `skill_category.py` | `backfill`, `infer`                                       | Write the closed-set `category:` field into skill/agent frontmatter. |
-| `kpi_dashboard.py`  | `render`, `summary`                                       | Emit a single Markdown dashboard joined across all quality sinks. |
-
-Asymmetric gates: downward transitions are automatic from a D-streak;
-upward transitions are deliberate. Archive needs aging (`active` for
-14 days in `_demoted`); delete needs a typed-slug confirmation even
-under `--auto`.
-
-## Category taxonomy
-
-Closed set: `framework`, `language`, `tool`, `pattern`, `workflow`,
-`meta`. The dashboard groups scores by category so you can see, e.g.,
-that your `framework`-tagged skills average a B but your
-`workflow`-tagged skills are mostly D — which tells you where to focus
-curation.
-
-Inference is precedence-ordered: `python + django` → `language`
-(language wins over framework). The backfill **never overwrites** an
-existing non-empty value — human edits win.
-
-Run once after the scorer has seeded the sidecars:
-
-```bash
-python -m skill_category backfill --dry-run
-python -m skill_category backfill            # apply
-```
-
-Unresolved slugs (no tag matched the taxonomy) are listed for manual
-curation.
-
-## Lifecycle CLI
-
-Four verbs. All are propose-and-confirm by default; `--auto` unlocks
-only the safe tiers (Watch + Demote).
-
-```bash
-# List every pending transition; no writes.
-python -m ctx_lifecycle review --dry-run
-
-# Apply all Watch/Demote transitions without prompting.
-python -m ctx_lifecycle review --auto
-
-# Archive a specific slug. Requires the demoted aging threshold to have passed.
-python -m ctx_lifecycle archive <slug>
-
-# Delete archived slugs that exceeded the delete threshold.
-# Requires typed-slug confirmation per entry even with --auto.
-python -m ctx_lifecycle purge
-
-# List archived slugs with optional diffs, or restore one.
-python -m ctx_lifecycle review-archived --show-diff
-python -m ctx_lifecycle review-archived --restore <slug>
-```
-
-Filesystem moves: demote → `<skills_dir>/_demoted/<slug>/`, archive →
-`<skills_dir>/_archive/<slug>/`. The scanner skips directories starting
-with `_`, so demoted/archived skills no longer show up to the router.
-
-Lifecycle state is persisted in a sibling sidecar at
-`~/.claude/skill-quality/<slug>.lifecycle.json`. Each transition is
-folded into the `history` array (capped at `history_max`) so you can
-audit how a slug ended up where it did.
-
-### The D-streak
-
-A skill needs `consecutive_d_to_demote` consecutive D-or-F grades to
-trigger a demote proposal. Any A/B/C grade resets the streak to 0. The
-default is 2 — one bad session is a blip, two in a row is a trend.
-
-## KPI dashboard
-
-Pure read-only. Walks both quality sidecars and lifecycle sidecars,
-joins them against `category:` in frontmatter (with inference
-fallback), and emits Markdown or JSON.
-
-```bash
-# Dump Markdown to stdout.
-python -m kpi_dashboard render
-
-# Persist to a file — good target for the cron / pre-push hook.
-python -m kpi_dashboard render --out ~/.claude/skill-quality/kpi.md
-
-# Machine-readable.
-python -m kpi_dashboard render --json --out kpi.json
-
-# Terse one-screen summary.
-python -m kpi_dashboard summary
-```
-
-### What's in the report
-
-- **Grade distribution** — A/B/C/D/F counts + percentages. Blank grade
-  (no score yet) rolls up to F so it surfaces in the "needs
-  attention" bucket.
-- **Lifecycle tiers** — counts across active/watch/demote/archive.
-- **Hard floors active** — how many slugs are failing intake or
-  never-loaded-stale.
-- **By category** — per-category count, average score of
-  scored entries, and a mini A/B/C/D/F mix table.
-- **Top demotion candidates** — up to `--limit N` (default 10) active
-  or watch-tier entries sorted by (D-streak desc, score asc). These
-  are the first slugs `ctx_lifecycle review --auto` will act on.
-- **Archived (restorable)** — every slug currently in the archive
-  tier; still recoverable via `review-archived --restore <slug>` until
-  `purge` deletes them.
-
-## Configuration
-
-All knobs live under `quality` in `src/config.json`:
-
-```json
-{
-  "quality": {
-    "lifecycle": {
-      "archive_threshold_days": 14.0,
-      "delete_threshold_days": 60.0,
-      "consecutive_d_to_demote": 2,
-      "demoted_subdir": "_demoted",
-      "archive_subdir": "_archive",
-      "history_max": 20
-    },
-    "dashboard": {
-      "default_top_n": 10,
-      "report_path": "~/.claude/skill-quality/kpi.md"
-    }
-  }
-}
-```
-
-Tighten `consecutive_d_to_demote` to 1 for aggressive pruning, or
-loosen `archive_threshold_days` if you want a longer grace window
-before a demoted skill gets archived.
-
-## Operational cadence
-
-A reasonable default rhythm, given the defaults above:
-
-- **Every session** — scorer hook runs automatically on session end.
-- **Weekly** — `ctx_lifecycle review --auto` to sweep Watch/Demote.
-- **Weekly** — `kpi_dashboard render --out …/kpi.md` for the digest.
-- **Monthly** — `ctx_lifecycle review` (no `--auto`) to surface
-  archive-ready demoted skills for manual approval.
-- **Quarterly** — `ctx_lifecycle purge` with typed-slug confirmation
-  to actually remove archived skills past the delete threshold.
-
-## Troubleshooting
-
-- **"unresolved" in backfill output** — the skill's tags don't match
-  the taxonomy. Either add a matching tag (e.g. `python`) or set
-  `category:` manually in frontmatter.
-- **Dashboard shows skills as F with no score** — they have a
-  lifecycle sidecar but no quality sidecar. That's intentional:
-  archived slugs whose quality sidecar was cleaned up still appear
-  in the tier and archive sections so you can restore them.
-- **`review --auto` refuses to archive or delete** — by design. Those
-  tiers require human approval.
+# Skill lifecycle & KPI dashboard — install & operations
+
+One page on running the lifecycle CLI, the category backfill, and the KPI
+dashboard. Prerequisite: the quality scorer is installed and has written at
+least one sidecar. See
+[skill-quality-install.md](./skill-quality-install.md).
+
+## What it does
+
+After the scorer has labeled everything, these three tools turn the
+labels into action:
+
+| Tool | CLI | Purpose |
+| ---- | --- | ------- |
+| `ctx_lifecycle.py`  | `review`, `demote`, `archive`, `purge`, `review-archived` | Move D/F-grade skills through `active → watch → demote → archive → deleted`. |
+| `skill_category.py` | `backfill`, `infer`                                       | Write the closed-set `category:` field into skill/agent frontmatter. |
+| `kpi_dashboard.py`  | `render`, `summary`                                       | Emit a single Markdown dashboard joined across all quality sinks. |
+
+Asymmetric gates: downward transitions are automatic from a D-streak;
+upward transitions are deliberate. Archive needs aging (`active` for
+14 days in `_demoted`); delete needs a typed-slug confirmation even
+under `--auto`.
+
+## Category taxonomy
+
+Closed set: `framework`, `language`, `tool`, `pattern`, `workflow`,
+`meta`. The dashboard groups scores by category so you can see, e.g.,
+that your `framework`-tagged skills average a B but your
+`workflow`-tagged skills are mostly D — which tells you where to focus
+curation.
+
+Inference is precedence-ordered: `python + django` → `language`
+(language wins over framework). The backfill **never overwrites** an
+existing non-empty value — human edits win.
+
+Run once after the scorer has seeded the sidecars:
+
+```bash
+python -m skill_category backfill --dry-run
+python -m skill_category backfill            # apply
+```
+
+Unresolved slugs (no tag matched the taxonomy) are listed for manual
+curation.
+
+## Lifecycle CLI
+
+Four verbs. All are propose-and-confirm by default; `--auto` unlocks
+only the safe tiers (Watch + Demote).
+
+```bash
+# List every pending transition; no writes.
+python -m ctx_lifecycle review --dry-run
+
+# Apply all Watch/Demote transitions without prompting.
+python -m ctx_lifecycle review --auto
+
+# Archive a specific slug. Requires the demoted aging threshold to have passed.
+python -m ctx_lifecycle archive <slug>
+
+# Delete archived slugs that exceeded the delete threshold.
+# Requires typed-slug confirmation per entry even with --auto.
+python -m ctx_lifecycle purge
+
+# List archived slugs with optional diffs, or restore one.
+python -m ctx_lifecycle review-archived --show-diff
+python -m ctx_lifecycle review-archived --restore <slug>
+```
+
+Filesystem moves: demote → `<skills_dir>/_demoted/<slug>/`, archive →
+`<skills_dir>/_archive/<slug>/`. The scanner skips directories starting
+with `_`, so demoted/archived skills no longer show up to the router.
+
+Lifecycle state is persisted in a sibling sidecar at
+`~/.claude/skill-quality/<slug>.lifecycle.json`. Each transition is
+folded into the `history` array (capped at `history_max`) so you can
+audit how a slug ended up where it did.
+
+### The D-streak
+
+A skill needs `consecutive_d_to_demote` consecutive D-or-F grades to
+trigger a demote proposal. Any A/B/C grade resets the streak to 0. The
+default is 2 — one bad session is a blip, two in a row is a trend.
+
+## KPI dashboard
+
+Pure read-only. Walks both quality sidecars and lifecycle sidecars,
+joins them against `category:` in frontmatter (with inference
+fallback), and emits Markdown or JSON.
+
+```bash
+# Dump Markdown to stdout.
+python -m kpi_dashboard render
+
+# Persist to a file — good target for the cron / pre-push hook.
+python -m kpi_dashboard render --out ~/.claude/skill-quality/kpi.md
+
+# Machine-readable.
+python -m kpi_dashboard render --json --out kpi.json
+
+# Terse one-screen summary.
+python -m kpi_dashboard summary
+```
+
+### What's in the report
+
+- **Grade distribution** — A/B/C/D/F counts + percentages. Blank grade
+  (no score yet) rolls up to F so it surfaces in the "needs
+  attention" bucket.
+- **Lifecycle tiers** — counts across active/watch/demote/archive.
+- **Hard floors active** — how many slugs are failing intake or
+  never-loaded-stale.
+- **By category** — per-category count, average score of
+  scored entries, and a mini A/B/C/D/F mix table.
+- **Top demotion candidates** — up to `--limit N` (default 10) active
+  or watch-tier entries sorted by (D-streak desc, score asc). These
+  are the first slugs `ctx_lifecycle review --auto` will act on.
+- **Archived (restorable)** — every slug currently in the archive
+  tier; still recoverable via `review-archived --restore <slug>` until
+  `purge` deletes them.
+
+## Configuration
+
+All knobs live under `quality` in `src/config.json`:
+
+```json
+{
+  "quality": {
+    "lifecycle": {
+      "archive_threshold_days": 14.0,
+      "delete_threshold_days": 60.0,
+      "consecutive_d_to_demote": 2,
+      "demoted_subdir": "_demoted",
+      "archive_subdir": "_archive",
+      "history_max": 20
+    },
+    "dashboard": {
+      "default_top_n": 10,
+      "report_path": "~/.claude/skill-quality/kpi.md"
+    }
+  }
+}
+```
+
+Tighten `consecutive_d_to_demote` to 1 for aggressive pruning, or
+loosen `archive_threshold_days` if you want a longer grace window
+before a demoted skill gets archived.
+
+## Operational cadence
+
+A reasonable default rhythm, given the defaults above:
+
+- **Every session** — scorer hook runs automatically on session end.
+- **Weekly** — `ctx_lifecycle review --auto` to sweep Watch/Demote.
+- **Weekly** — `kpi_dashboard render --out …/kpi.md` for the digest.
+- **Monthly** — `ctx_lifecycle review` (no `--auto`) to surface
+  archive-ready demoted skills for manual approval.
+- **Quarterly** — `ctx_lifecycle purge` with typed-slug confirmation
+  to actually remove archived skills past the delete threshold.
+
+## Troubleshooting
+
+- **"unresolved" in backfill output** — the skill's tags don't match
+  the taxonomy. Either add a matching tag (e.g. `python`) or set
+  `category:` manually in frontmatter.
+- **Dashboard shows skills as F with no score** — they have a
+  lifecycle sidecar but no quality sidecar. That's intentional:
+  archived slugs whose quality sidecar was cleaned up still appear
+  in the tier and archive sections so you can restore them.
+- **`review --auto` refuses to archive or delete** — by design. Those
+  tiers require human approval.

docs/skill-quality-install.md

@@ -1,170 +1,170 @@
-# Skill quality — install & operations
-
-One page on running the quality scorer: install the Stop hook, seed the
-sidecars, and verify the data flows into the wiki and the knowledge graph.
-
-## What it does
-
-Every installed skill and agent gets a continuous quality score in
-`[0.0, 1.0]` plus an A/B/C/D/F letter grade, derived from four signals:
-
-| Signal    | Weight | What it measures                                      |
-| --------- | -----: | ----------------------------------------------------- |
-| telemetry |   0.40 | Load count, recency, freshness in `skill-events.jsonl`|
-| intake    |   0.20 | Live re-run of the six install-time structural checks |
-| graph     |   0.25 | Degree + average edge weight in the wiki graph        |
-| routing   |   0.15 | Router hit-rate (neutral prior below 3 observations)  |
-
-Two hard floors override the weighted sum:
-
-- **`intake_fail`** — any structural check is currently failing → grade **F**.
-  Grade F is **only** produced by this hard floor; it is never returned by
-  the score-to-grade mapping function alone.  A score of 0.0 maps to **D**.
-- **`never_loaded_stale`** — no load events ever → grade capped at **D**.
-
-The score is mirrored to three on-disk sinks so every consumer sees the
-same number:
-
-1. `~/.claude/skill-quality/<slug>.json` — canonical machine-readable form.
-2. Wiki entity frontmatter — `quality_score`, `quality_grade`, `quality_updated_at`.
-3. Wiki body — a `## Quality` block between `<!-- quality:begin -->` markers.
-
-The knowledge-graph node attribute is a **separate consumer path**, not a
-write path from `persist_quality`: `wiki_graphify` reads the sidecar JSON
-produced by sink 1 and attaches `quality_score` / `quality_grade` to each
-node on its next build.
-
-## Register the Stop hook
-
-The hook runs once per session-end. It reads `skill-events.jsonl` since
-its last run, collects every slug that appeared, and calls
-`skill_quality.py recompute --slugs <comma-list>` — so scoring is
-incremental (touched skills only), not a full 2,000-page sweep.
-
-Edit `~/.claude/settings.json` and add, replacing `<REPO>` with the
-absolute path to this checkout (use forward slashes on Windows):
-
-```json
-{
-  "hooks": {
-    "Stop": [
-      {
-        "hooks": [
-          {
-            "type": "command",
-            "command": "python <REPO>/hooks/quality_on_session_end.py"
-          }
-        ]
-      }
-    ]
-  }
-}
-```
-
-The hook always exits 0: a scoring error will not block session
-shutdown.
-
-## Seed the sidecars (first run only)
-
-Run once after install so every installed skill has a baseline score:
-
-```bash
-ctx-skill-quality recompute --all
-```
-
-This walks `~/.claude/skills/*/SKILL.md` and `~/.claude/agents/*.md`,
-scores each, and writes the three on-disk sinks. Expect ~15–30s depending on
-corpus size and disk.
-
-## CLI reference
-
-```bash
-# Full recompute (use sparingly; the Stop hook handles incrementals).
-ctx-skill-quality recompute --all
-
-# One slug.
-ctx-skill-quality recompute --slug python-testing
-
-# Show the most recent score.
-ctx-skill-quality show python-testing
-
-# Signal-by-signal breakdown with evidence.
-ctx-skill-quality explain python-testing
-
-# List every slug with its grade, filtered.
-ctx-skill-quality list --grade D
-```
-
-All verbs accept `--json` for piping into other tools.
-
-## Graph integration
-
-`wiki_graphify.py` reads the sidecar directory automatically and
-attaches `quality_score` and `quality_grade` to every matching node. The
-Obsidian graph view can then color nodes by grade — configure the
-`quality_grade` property in Obsidian's graph settings.
-
-Nodes without a sidecar get `quality_score: null` and `quality_grade:
-null` so downstream consumers can always read the attribute safely.
-
-## Configuration
-
-All knobs live in `src/config.json` under the top-level `quality` key:
-
-```json
-{
-  "quality": {
-    "weights": {
-      "telemetry": 0.40, "intake": 0.20, "graph": 0.25, "routing": 0.15
-    },
-    "agent_weights": {
-      "telemetry": 0.15, "intake": 0.30, "graph": 0.35, "routing": 0.20
-    },
-    "grade_thresholds": {"A": 0.80, "B": 0.60, "C": 0.40},
-    "stale_threshold_days": 30.0,
-    "recent_window_days": 14.0,
-    "min_body_chars": 120,
-    "paths": {
-      "sidecar_dir": "~/.claude/skill-quality",
-      "router_trace": "~/.claude/router-trace.jsonl"
-    }
-  }
-}
-```
-
-`ctx_config.Config` exposes this through `cfg.get("quality", {})`. User
-overrides in `~/.claude/skill-system-config.json` deep-merge over the
-repo defaults, so you can pin only the keys you want to change.
-
-Both weight vectors must sum to 1.0 (±0.01) and grade thresholds must
-satisfy `0 ≤ C ≤ B ≤ A ≤ 1` — `QualityConfig.__post_init__` will raise
-on bad values, catching typos before they pollute sidecars.
-
-### Why two weight vectors
-
-Skills and agents differ in how they're invoked. Skills are loaded
-automatically by the router, so **telemetry** (load counts, recency) is
-the strongest post-install quality signal — hence 0.40 weight.
-
-Agents are invoked via the Agent tool, deliberately and rarely. A
-seldom-used agent isn't stale, it's specialized. The agent weights
-shift mass onto **graph connectedness** (0.35) and **intake structure**
-(0.30) so agents aren't penalized for having an empty telemetry stream.
-The `never_loaded_stale` hard floor, which caps skills at D when they
-have zero load events, does **not** apply to agents for the same
-reason.
-
-## Troubleshooting
-
-- **Every skill grades D.** Telemetry hasn't accumulated enough load
-  events yet. This is expected on a fresh install; the stop-hook will
-  pick up real usage over the next few sessions.
-- **A recently-edited skill now grades F.** Open the sidecar and look
-  at `signals.intake.evidence.checks` — one of the six structural
-  checks is failing. Fix the file and rerun `recompute --slug <name>`.
-- **Wiki page has two `## Quality` sections.** Shouldn't happen —
-  `persist_quality` is idempotent via the HTML-comment markers. If it
-  does, delete both blocks and rerun `recompute`; the first pass will
-  re-emit exactly one.
-- **Graph view shows no color.** Run `ctx-wiki-graphify
-  --graph-only` to rebuild; it reads sidecars fresh on every build.
+# Skill quality — install & operations
+
+One page on running the quality scorer: install the Stop hook, seed the
+sidecars, and verify the data flows into the wiki and the knowledge graph.
+
+## What it does
+
+Every installed skill and agent gets a continuous quality score in
+`[0.0, 1.0]` plus an A/B/C/D/F letter grade, derived from four signals:
+
+| Signal    | Weight | What it measures                                      |
+| --------- | -----: | ----------------------------------------------------- |
+| telemetry |   0.40 | Load count, recency, freshness in `skill-events.jsonl`|
+| intake    |   0.20 | Live re-run of the six install-time structural checks |
+| graph     |   0.25 | Degree + average edge weight in the wiki graph        |
+| routing   |   0.15 | Router hit-rate (neutral prior below 3 observations)  |
+
+Two hard floors override the weighted sum:
+
+- **`intake_fail`** — any structural check is currently failing → grade **F**.
+  Grade F is **only** produced by this hard floor; it is never returned by
+  the score-to-grade mapping function alone.  A score of 0.0 maps to **D**.
+- **`never_loaded_stale`** — no load events ever → grade capped at **D**.
+
+The score is mirrored to three on-disk sinks so every consumer sees the
+same number:
+
+1. `~/.claude/skill-quality/<slug>.json` — canonical machine-readable form.
+2. Wiki entity frontmatter — `quality_score`, `quality_grade`, `quality_updated_at`.
+3. Wiki body — a `## Quality` block between `<!-- quality:begin -->` markers.
+
+The knowledge-graph node attribute is a **separate consumer path**, not a
+write path from `persist_quality`: `wiki_graphify` reads the sidecar JSON
+produced by sink 1 and attaches `quality_score` / `quality_grade` to each
+node on its next build.
+
+## Register the Stop hook
+
+The hook runs once per session-end. It reads `skill-events.jsonl` since
+its last run, collects every slug that appeared, and calls
+`skill_quality.py recompute --slugs <comma-list>` — so scoring is
+incremental (touched skills only), not a full 2,000-page sweep.
+
+Edit `~/.claude/settings.json` and add, replacing `<REPO>` with the
+absolute path to this checkout (use forward slashes on Windows):
+
+```json
+{
+  "hooks": {
+    "Stop": [
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "python <REPO>/hooks/quality_on_session_end.py"
+          }
+        ]
+      }
+    ]
+  }
+}
+```
+
+The hook always exits 0: a scoring error will not block session
+shutdown.
+
+## Seed the sidecars (first run only)
+
+Run once after install so every installed skill has a baseline score:
+
+```bash
+ctx-skill-quality recompute --all
+```
+
+This walks `~/.claude/skills/*/SKILL.md` and `~/.claude/agents/*.md`,
+scores each, and writes the three on-disk sinks. Expect ~15–30s depending on
+corpus size and disk.
+
+## CLI reference
+
+```bash
+# Full recompute (use sparingly; the Stop hook handles incrementals).
+ctx-skill-quality recompute --all
+
+# One slug.
+ctx-skill-quality recompute --slug python-testing
+
+# Show the most recent score.
+ctx-skill-quality show python-testing
+
+# Signal-by-signal breakdown with evidence.
+ctx-skill-quality explain python-testing
+
+# List every slug with its grade, filtered.
+ctx-skill-quality list --grade D
+```
+
+All verbs accept `--json` for piping into other tools.
+
+## Graph integration
+
+`wiki_graphify.py` reads the sidecar directory automatically and
+attaches `quality_score` and `quality_grade` to every matching node. The
+Obsidian graph view can then color nodes by grade — configure the
+`quality_grade` property in Obsidian's graph settings.
+
+Nodes without a sidecar get `quality_score: null` and `quality_grade:
+null` so downstream consumers can always read the attribute safely.
+
+## Configuration
+
+All knobs live in `src/config.json` under the top-level `quality` key:
+
+```json
+{
+  "quality": {
+    "weights": {
+      "telemetry": 0.40, "intake": 0.20, "graph": 0.25, "routing": 0.15
+    },
+    "agent_weights": {
+      "telemetry": 0.15, "intake": 0.30, "graph": 0.35, "routing": 0.20
+    },
+    "grade_thresholds": {"A": 0.80, "B": 0.60, "C": 0.40},
+    "stale_threshold_days": 30.0,
+    "recent_window_days": 14.0,
+    "min_body_chars": 120,
+    "paths": {
+      "sidecar_dir": "~/.claude/skill-quality",
+      "router_trace": "~/.claude/router-trace.jsonl"
+    }
+  }
+}
+```
+
+`ctx_config.Config` exposes this through `cfg.get("quality", {})`. User
+overrides in `~/.claude/skill-system-config.json` deep-merge over the
+repo defaults, so you can pin only the keys you want to change.
+
+Both weight vectors must sum to 1.0 (±0.01) and grade thresholds must
+satisfy `0 ≤ C ≤ B ≤ A ≤ 1` — `QualityConfig.__post_init__` will raise
+on bad values, catching typos before they pollute sidecars.
+
+### Why two weight vectors
+
+Skills and agents differ in how they're invoked. Skills are loaded
+automatically by the router, so **telemetry** (load counts, recency) is
+the strongest post-install quality signal — hence 0.40 weight.
+
+Agents are invoked via the Agent tool, deliberately and rarely. A
+seldom-used agent isn't stale, it's specialized. The agent weights
+shift mass onto **graph connectedness** (0.35) and **intake structure**
+(0.30) so agents aren't penalized for having an empty telemetry stream.
+The `never_loaded_stale` hard floor, which caps skills at D when they
+have zero load events, does **not** apply to agents for the same
+reason.
+
+## Troubleshooting
+
+- **Every skill grades D.** Telemetry hasn't accumulated enough load
+  events yet. This is expected on a fresh install; the stop-hook will
+  pick up real usage over the next few sessions.
+- **A recently-edited skill now grades F.** Open the sidecar and look
+  at `signals.intake.evidence.checks` — one of the six structural
+  checks is failing. Fix the file and rerun `recompute --slug <name>`.
+- **Wiki page has two `## Quality` sections.** Shouldn't happen —
+  `persist_quality` is idempotent via the HTML-comment markers. If it
+  does, delete both blocks and rerun `recompute`; the first pass will
+  re-emit exactly one.
+- **Graph view shows no color.** Run `ctx-wiki-graphify
+  --graph-only` to rebuild; it reads sidecars fresh on every build.

docs/skill-router/index.md

@@ -1,51 +1,51 @@
-# Skill router
-
-The skill router decides which skills, agents, and MCP servers are useful for
-the active repository and current development task. Harnesses are recommended in
-the custom-model onboarding flow; after a harness is attached, the same capped
-skills/agents/MCP recommendation layer can be used by that host.
-
-## Problem
-
-Every loaded skill, agent, and MCP server costs tokens, attention, and tool
-surface area. Most sessions need a small top-scored bundle from the 91K+ skills,
-460+ agents, and 10K+ MCP servers in the shipped graph, not the whole inventory.
-Loading too much:
-
-- wastes context on irrelevant instructions,
-- causes the wrong helper to trigger for a task,
-- slows response time, and
-- creates conflicting instructions between helpers.
-
-## Architecture
-
-```text
-ctx/
-|-- src/scan_repo.py                         # Repo scanner -> stack profile
-|-- src/ctx/core/resolve/resolve_skills.py   # Profile -> load/unload manifest
-|-- src/ctx/core/resolve/recommendations.py  # Shared recommendation engine
-|-- src/ctx/adapters/                        # Claude Code hooks + generic tools
-`-- graph/wiki-graph-runtime.tar.gz          # Shipped graph/wiki runtime
-```
-
-## Flow
-
-1. `ctx-scan-repo --repo . --recommend` scans the repository and produces stack signals.
-2. The shared resolver scores graph/wiki entities by tags, categories, semantic
-   edges, usage, quality, and configured gates.
-3. The resolver returns a capped manifest: what to load, what to unload, and why.
-4. The user confirms load/unload changes unless they configured automatic mode.
-5. Usage and quality signals are recorded so future recommendations improve.
-
-The same recommender is used by the CLI, MCP/library tools, Claude Code hooks,
-and attached harness hosts. Entry points should differ only in transport and
-confirmation UX, not in ranking logic.
-
-## Reference Pages
-
-- [Stack signatures](../stack-signatures.md) - file/config patterns used to
-  identify stack signals.
-- [Skill-stack matrix](../skill-stack-matrix.md) - stack-to-capability mapping
-  used as scanner evidence.
-- [Entity source registry](../marketplace-registry.md) - skill, GitHub,
-  MCP, harness, and local sources plus update rules.
+# Skill router
+
+The skill router decides which skills, agents, and MCP servers are useful for
+the active repository and current development task. Harnesses are recommended in
+the custom-model onboarding flow; after a harness is attached, the same capped
+skills/agents/MCP recommendation layer can be used by that host.
+
+## Problem
+
+Every loaded skill, agent, and MCP server costs tokens, attention, and tool
+surface area. Most sessions need a small top-scored bundle from the 91K+ skills,
+460+ agents, and 10K+ MCP servers in the shipped graph, not the whole inventory.
+Loading too much:
+
+- wastes context on irrelevant instructions,
+- causes the wrong helper to trigger for a task,
+- slows response time, and
+- creates conflicting instructions between helpers.
+
+## Architecture
+
+```text
+ctx/
+|-- src/scan_repo.py                         # Repo scanner -> stack profile
+|-- src/ctx/core/resolve/resolve_skills.py   # Profile -> load/unload manifest
+|-- src/ctx/core/resolve/recommendations.py  # Shared recommendation engine
+|-- src/ctx/adapters/                        # Claude Code hooks + generic tools
+`-- graph/wiki-graph-runtime.tar.gz          # Shipped graph/wiki runtime
+```
+
+## Flow
+
+1. `ctx-scan-repo --repo . --recommend` scans the repository and produces stack signals.
+2. The shared resolver scores graph/wiki entities by tags, categories, semantic
+   edges, usage, quality, and configured gates.
+3. The resolver returns a capped manifest: what to load, what to unload, and why.
+4. The user confirms load/unload changes unless they configured automatic mode.
+5. Usage and quality signals are recorded so future recommendations improve.
+
+The same recommender is used by the CLI, MCP/library tools, Claude Code hooks,
+and attached harness hosts. Entry points should differ only in transport and
+confirmation UX, not in ranking logic.
+
+## Reference Pages
+
+- [Stack signatures](../stack-signatures.md) - file/config patterns used to
+  identify stack signals.
+- [Skill-stack matrix](../skill-stack-matrix.md) - stack-to-capability mapping
+  used as scanner evidence.
+- [Entity source registry](../marketplace-registry.md) - skill, GitHub,
+  MCP, harness, and local sources plus update rules.

docs/skill-stack-matrix.md

@@ -1,166 +1,166 @@
-# Skill-Stack Matrix
-
-> Maps stack identifiers to the skills that serve them.
-> Used as resolver evidence; final ranking still goes through the shared
-> recommendation engine and graph/wiki scores.
-
-## Table of Contents
-1. [Matrix Format](#matrix-format)
-2. [The Matrix](#the-matrix)
-3. [Companion Rules](#companion-rules)
-4. [Conflict Rules](#conflict-rules)
-
----
-
-## Matrix Format
-
-Each entry:
-- **Stack IDs**: which detected stacks trigger this skill
-- **Skill**: skill/entity slug in the graph/wiki catalog
-- **Priority Base**: starting priority before signal boosts
-- **Required**: must-load if stack detected, vs nice-to-have
-- **Companions**: skills that should co-load
-- **Conflicts**: skills that should not co-load
-
-## The Matrix
-
-### Document Creation Skills
-
-| Skill | Stack IDs | Priority | Required | Path Pattern |
-|-------|-----------|----------|----------|--------------|
-| docx | (any -- triggered by user request) | 2 | no | graph/wiki or local skill path |
-| pdf | (any -- triggered by user request) | 2 | no | graph/wiki or local skill path |
-| pptx | (any -- triggered by user request) | 2 | no | graph/wiki or local skill path |
-| xlsx | (any -- triggered by user request) | 2 | no | graph/wiki or local skill path |
-
-> Note: document skills are demand-loaded, not stack-loaded. They activate on user
-> request ("make a presentation") not on repo content. The router keeps them in a
-> "standby" pool -- not loaded into context, but available for instant load.
-
-### Frontend Skills
-
-| Skill | Stack IDs | Priority | Required |
-|-------|-----------|----------|----------|
-| frontend-design | react, vue, angular, svelte, nextjs, nuxt, html, css | 8 | yes |
-| react | react, nextjs | 7 | yes |
-| vue | vue, nuxt | 7 | yes |
-| angular | angular | 7 | yes |
-| svelte | svelte | 7 | yes |
-| tailwind | tailwindcss | 5 | no |
-| css-modules | css-modules | 4 | no |
-
-### Backend Skills
-
-| Skill | Stack IDs | Priority | Required |
-|-------|-----------|----------|----------|
-| fastapi | fastapi | 8 | yes |
-| django | django | 8 | yes |
-| flask | flask | 7 | yes |
-| express | express | 8 | yes |
-| nestjs | nestjs | 8 | yes |
-| rails | rails | 8 | yes |
-| gin | gin | 7 | yes |
-| actix | actix | 7 | yes |
-
-### Data Skills
-
-| Skill | Stack IDs | Priority | Required |
-|-------|-----------|----------|----------|
-| sqlalchemy | sqlalchemy, alembic | 6 | yes |
-| prisma | prisma | 6 | yes |
-| typeorm | typeorm | 6 | yes |
-| drizzle | drizzle | 6 | yes |
-| redis | redis | 4 | no |
-| kafka | kafka | 5 | no |
-| dbt | dbt | 6 | yes |
-
-### Infrastructure Skills
-
-| Skill | Stack IDs | Priority | Required |
-|-------|-----------|----------|----------|
-| docker | docker, docker-compose | 6 | yes |
-| kubernetes | kubernetes, helm, kustomize | 6 | yes |
-| terraform | terraform | 7 | yes |
-| github-actions | github-actions | 5 | yes |
-| gitlab-ci | gitlab-ci | 5 | yes |
-| aws | aws-cdk, aws-sam | 7 | yes |
-| vercel | vercel | 4 | no |
-
-### AI/Agent Skills
-
-| Skill | Stack IDs | Priority | Required |
-|-------|-----------|----------|----------|
-| langchain | langchain | 7 | yes |
-| llamaindex | llamaindex | 7 | yes |
-| mcp-dev | mcp | 7 | yes |
-| pytorch | pytorch | 6 | yes |
-| huggingface | huggingface | 6 | yes |
-| openai-sdk | openai-sdk | 5 | no |
-| anthropic-sdk | anthropic-sdk | 5 | no |
-
-### Quality Skills
-
-| Skill | Stack IDs | Priority | Required |
-|-------|-----------|----------|----------|
-| pytest | pytest | 5 | yes |
-| jest | jest, vitest | 5 | yes |
-| cypress | cypress | 4 | no |
-| playwright | playwright | 4 | no |
-| eslint | eslint | 3 | no |
-| ruff | ruff | 3 | no |
-
-### Documentation Skills
-
-| Skill | Stack IDs | Priority | Required |
-|-------|-----------|----------|----------|
-| openapi | openapi | 5 | yes |
-| graphql | graphql | 5 | yes |
-| mkdocs | mkdocs | 4 | no |
-| docusaurus | docusaurus | 4 | no |
-
-### Meta Skills (always available, never unloaded)
-
-| Skill | Stack IDs | Priority | Required | Notes |
-|-------|-----------|----------|----------|-------|
-| skill-router | * | 99 | yes | This skill -- always loaded |
-| file-reading | * | 50 | yes | Core capability |
-| skill-creator | * | 10 | no | Standby pool |
-| product-self-knowledge | * | 10 | no | Standby pool |
-
----
-
-## Companion Rules
-
-When skill A is loaded, also load skill B if its stack is detected:
-
-| Primary Skill | Companion | Condition |
-|---------------|-----------|-----------|
-| fastapi | sqlalchemy | DB migrations detected |
-| fastapi | openapi | OpenAPI spec file exists |
-| django | django-orm | (always with django) |
-| react | tailwind | tailwind.config.* exists |
-| docker | docker-compose | docker-compose.yml exists |
-| kubernetes | helm | Chart.yaml exists |
-| terraform | aws | provider "aws" in *.tf |
-| langchain | openai-sdk | openai in deps |
-| pytest | coverage | .coveragerc or coverage config exists |
-
-## Conflict Rules
-
-These skills should not be co-loaded (pick the one with higher confidence/priority):
-
-| Skill A | Skill B | Resolution |
-|---------|---------|------------|
-| flask | fastapi | Higher confidence wins |
-| flask | django | Higher confidence wins |
-| jest | vitest | Higher confidence wins |
-| webpack | vite | Higher confidence wins |
-| npm | yarn | Check lock file |
-| npm | pnpm | Check lock file |
-| yarn | pnpm | Check lock file |
-| react | vue | Both can coexist in monorepo |
-| sqlalchemy | prisma | Both can coexist if different services |
-
-> Conflict resolution: check if the repo is a monorepo. In monorepos, "conflicting"
-> skills may serve different packages and should both load. In single-package repos,
-> pick the one with higher confidence.
+# Skill-Stack Matrix
+
+> Maps stack identifiers to the skills that serve them.
+> Used as resolver evidence; final ranking still goes through the shared
+> recommendation engine and graph/wiki scores.
+
+## Table of Contents
+1. [Matrix Format](#matrix-format)
+2. [The Matrix](#the-matrix)
+3. [Companion Rules](#companion-rules)
+4. [Conflict Rules](#conflict-rules)
+
+---
+
+## Matrix Format
+
+Each entry:
+- **Stack IDs**: which detected stacks trigger this skill
+- **Skill**: skill/entity slug in the graph/wiki catalog
+- **Priority Base**: starting priority before signal boosts
+- **Required**: must-load if stack detected, vs nice-to-have
+- **Companions**: skills that should co-load
+- **Conflicts**: skills that should not co-load
+
+## The Matrix
+
+### Document Creation Skills
+
+| Skill | Stack IDs | Priority | Required | Path Pattern |
+|-------|-----------|----------|----------|--------------|
+| docx | (any -- triggered by user request) | 2 | no | graph/wiki or local skill path |
+| pdf | (any -- triggered by user request) | 2 | no | graph/wiki or local skill path |
+| pptx | (any -- triggered by user request) | 2 | no | graph/wiki or local skill path |
+| xlsx | (any -- triggered by user request) | 2 | no | graph/wiki or local skill path |
+
+> Note: document skills are demand-loaded, not stack-loaded. They activate on user
+> request ("make a presentation") not on repo content. The router keeps them in a
+> "standby" pool -- not loaded into context, but available for instant load.
+
+### Frontend Skills
+
+| Skill | Stack IDs | Priority | Required |
+|-------|-----------|----------|----------|
+| frontend-design | react, vue, angular, svelte, nextjs, nuxt, html, css | 8 | yes |
+| react | react, nextjs | 7 | yes |
+| vue | vue, nuxt | 7 | yes |
+| angular | angular | 7 | yes |
+| svelte | svelte | 7 | yes |
+| tailwind | tailwindcss | 5 | no |
+| css-modules | css-modules | 4 | no |
+
+### Backend Skills
+
+| Skill | Stack IDs | Priority | Required |
+|-------|-----------|----------|----------|
+| fastapi | fastapi | 8 | yes |
+| django | django | 8 | yes |
+| flask | flask | 7 | yes |
+| express | express | 8 | yes |
+| nestjs | nestjs | 8 | yes |
+| rails | rails | 8 | yes |
+| gin | gin | 7 | yes |
+| actix | actix | 7 | yes |
+
+### Data Skills
+
+| Skill | Stack IDs | Priority | Required |
+|-------|-----------|----------|----------|
+| sqlalchemy | sqlalchemy, alembic | 6 | yes |
+| prisma | prisma | 6 | yes |
+| typeorm | typeorm | 6 | yes |
+| drizzle | drizzle | 6 | yes |
+| redis | redis | 4 | no |
+| kafka | kafka | 5 | no |
+| dbt | dbt | 6 | yes |
+
+### Infrastructure Skills
+
+| Skill | Stack IDs | Priority | Required |
+|-------|-----------|----------|----------|
+| docker | docker, docker-compose | 6 | yes |
+| kubernetes | kubernetes, helm, kustomize | 6 | yes |
+| terraform | terraform | 7 | yes |
+| github-actions | github-actions | 5 | yes |
+| gitlab-ci | gitlab-ci | 5 | yes |
+| aws | aws-cdk, aws-sam | 7 | yes |
+| vercel | vercel | 4 | no |
+
+### AI/Agent Skills
+
+| Skill | Stack IDs | Priority | Required |
+|-------|-----------|----------|----------|
+| langchain | langchain | 7 | yes |
+| llamaindex | llamaindex | 7 | yes |
+| mcp-dev | mcp | 7 | yes |
+| pytorch | pytorch | 6 | yes |
+| huggingface | huggingface | 6 | yes |
+| openai-sdk | openai-sdk | 5 | no |
+| anthropic-sdk | anthropic-sdk | 5 | no |
+
+### Quality Skills
+
+| Skill | Stack IDs | Priority | Required |
+|-------|-----------|----------|----------|
+| pytest | pytest | 5 | yes |
+| jest | jest, vitest | 5 | yes |
+| cypress | cypress | 4 | no |
+| playwright | playwright | 4 | no |
+| eslint | eslint | 3 | no |
+| ruff | ruff | 3 | no |
+
+### Documentation Skills
+
+| Skill | Stack IDs | Priority | Required |
+|-------|-----------|----------|----------|
+| openapi | openapi | 5 | yes |
+| graphql | graphql | 5 | yes |
+| mkdocs | mkdocs | 4 | no |
+| docusaurus | docusaurus | 4 | no |
+
+### Meta Skills (always available, never unloaded)
+
+| Skill | Stack IDs | Priority | Required | Notes |
+|-------|-----------|----------|----------|-------|
+| skill-router | * | 99 | yes | This skill -- always loaded |
+| file-reading | * | 50 | yes | Core capability |
+| skill-creator | * | 10 | no | Standby pool |
+| product-self-knowledge | * | 10 | no | Standby pool |
+
+---
+
+## Companion Rules
+
+When skill A is loaded, also load skill B if its stack is detected:
+
+| Primary Skill | Companion | Condition |
+|---------------|-----------|-----------|
+| fastapi | sqlalchemy | DB migrations detected |
+| fastapi | openapi | OpenAPI spec file exists |
+| django | django-orm | (always with django) |
+| react | tailwind | tailwind.config.* exists |
+| docker | docker-compose | docker-compose.yml exists |
+| kubernetes | helm | Chart.yaml exists |
+| terraform | aws | provider "aws" in *.tf |
+| langchain | openai-sdk | openai in deps |
+| pytest | coverage | .coveragerc or coverage config exists |
+
+## Conflict Rules
+
+These skills should not be co-loaded (pick the one with higher confidence/priority):
+
+| Skill A | Skill B | Resolution |
+|---------|---------|------------|
+| flask | fastapi | Higher confidence wins |
+| flask | django | Higher confidence wins |
+| jest | vitest | Higher confidence wins |
+| webpack | vite | Higher confidence wins |
+| npm | yarn | Check lock file |
+| npm | pnpm | Check lock file |
+| yarn | pnpm | Check lock file |
+| react | vue | Both can coexist in monorepo |
+| sqlalchemy | prisma | Both can coexist if different services |
+
+> Conflict resolution: check if the repo is a monorepo. In monorepos, "conflicting"
+> skills may serve different packages and should both load. In single-package repos,
+> pick the one with higher confidence.

docs/skills-health.md

@@ -1,108 +1,108 @@
-# Skill health dashboard
-
-[`src/skill_health.py`](https://github.com/stevesolun/ctx/blob/main/src/skill_health.py)
-scans `~/.claude/skills/` and `~/.claude/agents/` for structural and
-catalog issues, then produces a JSON or human-readable dashboard. It
-also self-heals catalog drift — without ever modifying a SKILL.md.
-
-## What it checks
-
-For each skill (`~/.claude/skills/<name>/SKILL.md`) and each agent
-(`~/.claude/agents/<name>.md`):
-
-| Code | Severity | Condition |
-|---|---|---|
-| `missing-file` | error | Skill directory has no SKILL.md |
-| `unreadable` | error | File exists but can't be decoded as UTF-8 |
-| `no-frontmatter` | error | Missing or malformed `---` YAML fence |
-| `frontmatter-missing-name` | error | Frontmatter has no `name:` field |
-| `frontmatter-missing-description` | warning | Missing `description:` (router relevance suffers) |
-| `empty-body` | error | Fewer than `min_body_lines` non-blank lines |
+# Skill health dashboard
+
+[`src/skill_health.py`](https://github.com/stevesolun/ctx/blob/main/src/skill_health.py)
+scans `~/.claude/skills/` and `~/.claude/agents/` for structural and
+catalog issues, then produces a JSON or human-readable dashboard. It
+also self-heals catalog drift — without ever modifying a SKILL.md.
+
+## What it checks
+
+For each skill (`~/.claude/skills/<name>/SKILL.md`) and each agent
+(`~/.claude/agents/<name>.md`):
+
+| Code | Severity | Condition |
+|---|---|---|
+| `missing-file` | error | Skill directory has no SKILL.md |
+| `unreadable` | error | File exists but can't be decoded as UTF-8 |
+| `no-frontmatter` | error | Missing or malformed `---` YAML fence |
+| `frontmatter-missing-name` | error | Frontmatter has no `name:` field |
+| `frontmatter-missing-description` | warning | Missing `description:` (router relevance suffers) |
+| `empty-body` | error | Fewer than `min_body_lines` non-blank lines |
 | `over-threshold` | warning | Line count exceeds `skill_transformer.line_threshold`; the packaged default is 180 and users can override it in `~/.claude/skill-system-config.json` |
-
-## Drift detection
-
-`DriftReport` cross-references three sources:
-
-- on-disk entities (skills + agents),
-- `~/.claude/skill-manifest.json` → `load[].skill` entries,
-- `~/.claude/pending-skills.json` → `graph_suggestions[].name` and
-  `unmatched_signals[]`.
-
-Anything in the manifest or pending file that doesn't exist on disk
-becomes an *orphan*. Orphans are the only thing `heal` is allowed to
-touch.
-
-## Self-healing
-
-```bash
-ctx-skill-health heal
-```
-
-- drops orphaned entries from `skill-manifest.json`
-- drops orphaned entries from `pending-skills.json`
-- writes atomically (`tempfile.mkstemp` + `os.replace`)
-- never modifies SKILL.md files or agent .md files
-
-If nothing needs healing, prints `[heal] nothing to do.` and exits 0.
-
-## CLI
-
-```bash
-# Emit a full JSON report
-ctx-skill-health scan
-
-# Pretty dashboard
-ctx-skill-health dashboard
-
-# CI gate: exit 2 if any error-severity issue or drift is present
-ctx-skill-health check --strict
-
-# Apply safe autofixes to manifest + pending
-ctx-skill-health heal
-```
-
-## Data model
-
-```python
-@dataclass(frozen=True)
-class Issue:
-    code: str
-    severity: str      # "warning" | "error"
-    message: str
-
-@dataclass(frozen=True)
-class EntityHealth:
-    name: str
-    kind: str          # "skill" | "agent"
-    path: str
-    lines: int
-    has_frontmatter: bool
-    issues: tuple[Issue, ...] = ()
-
-@dataclass(frozen=True)
-class DriftReport:
-    orphaned_manifest: tuple[str, ...] = ()
-    orphaned_pending: tuple[str, ...] = ()
-
-@dataclass(frozen=True)
-class HealthReport:
-    generated_at: float
-    entities: tuple[EntityHealth, ...]
-    drift: DriftReport
-    totals: dict[str, int]
-```
-
-`HealthReport.has_errors` is true when any entity has severity `error`
-*or* drift is non-empty — that's the single predicate behind
-`check --strict`'s exit code.
-
-## Related
-
-- [Memory anchoring](memory-anchor.md) — dead-reference detection for
-  auto-memory notes.
-- `src/skill_quality.py` (v0.5.0+) — the four-signal quality scorer
-  (telemetry 0.40, intake 0.20, graph 0.25, routing 0.15) that writes
-  per-entity sidecars and surfaces A/B/C/D/F grades. The `skill_health`
-  CLI above focuses on *structural* correctness; `skill_quality`
-  focuses on *behavioral* quality over time.
+
+## Drift detection
+
+`DriftReport` cross-references three sources:
+
+- on-disk entities (skills + agents),
+- `~/.claude/skill-manifest.json` → `load[].skill` entries,
+- `~/.claude/pending-skills.json` → `graph_suggestions[].name` and
+  `unmatched_signals[]`.
+
+Anything in the manifest or pending file that doesn't exist on disk
+becomes an *orphan*. Orphans are the only thing `heal` is allowed to
+touch.
+
+## Self-healing
+
+```bash
+ctx-skill-health heal
+```
+
+- drops orphaned entries from `skill-manifest.json`
+- drops orphaned entries from `pending-skills.json`
+- writes atomically (`tempfile.mkstemp` + `os.replace`)
+- never modifies SKILL.md files or agent .md files
+
+If nothing needs healing, prints `[heal] nothing to do.` and exits 0.
+
+## CLI
+
+```bash
+# Emit a full JSON report
+ctx-skill-health scan
+
+# Pretty dashboard
+ctx-skill-health dashboard
+
+# CI gate: exit 2 if any error-severity issue or drift is present
+ctx-skill-health check --strict
+
+# Apply safe autofixes to manifest + pending
+ctx-skill-health heal
+```
+
+## Data model
+
+```python
+@dataclass(frozen=True)
+class Issue:
+    code: str
+    severity: str      # "warning" | "error"
+    message: str
+
+@dataclass(frozen=True)
+class EntityHealth:
+    name: str
+    kind: str          # "skill" | "agent"
+    path: str
+    lines: int
+    has_frontmatter: bool
+    issues: tuple[Issue, ...] = ()
+
+@dataclass(frozen=True)
+class DriftReport:
+    orphaned_manifest: tuple[str, ...] = ()
+    orphaned_pending: tuple[str, ...] = ()
+
+@dataclass(frozen=True)
+class HealthReport:
+    generated_at: float
+    entities: tuple[EntityHealth, ...]
+    drift: DriftReport
+    totals: dict[str, int]
+```
+
+`HealthReport.has_errors` is true when any entity has severity `error`
+*or* drift is non-empty — that's the single predicate behind
+`check --strict`'s exit code.
+
+## Related
+
+- [Memory anchoring](memory-anchor.md) — dead-reference detection for
+  auto-memory notes.
+- `src/skill_quality.py` (v0.5.0+) — the four-signal quality scorer
+  (telemetry 0.40, intake 0.20, graph 0.25, routing 0.15) that writes
+  per-entity sidecars and surfaces A/B/C/D/F grades. The `skill_health`
+  CLI above focuses on *structural* correctness; `skill_quality`
+  focuses on *behavioral* quality over time.

docs/toolbox/behavior-miner.md

@@ -1,90 +1,90 @@
-# Behavior miner
-
-[`src/behavior_miner.py`](https://github.com/stevesolun/ctx/blob/main/src/behavior_miner.py)
-watches your invocation patterns and proposes toolbox tweaks grounded in
-real evidence.
-
-## What it collects
-
-Four signal families, each with `MIN_EVIDENCE = 3` before a suggestion
-can surface:
-
-| Signal | Source | Example suggestion |
-|---|---|---|
-| **Co-invocation** | Pairs of agents invoked in the same session | "You ran `code-reviewer` + `security-reviewer` together 4 times — consider a bundle." |
-| **Skill cadence** | Skill load frequency over time | "`python-patterns` loaded every session — promote to `pre`." |
-| **File-type** | File extensions of work-in-progress | "60% of your diffs touch `.tf` files — consider a Terraform toolbox." |
-| **Commit-type** | Conventional Commit parsing | "8 of your last 10 commits are `fix:` — consider a pre-commit test toolbox." |
-
-## User profile
-
-Signals aggregate into `~/.claude/user-profile.json`:
-
-```jsonc
-{
-  "version": 1,
-  "updated_at": 1713456789,
-  "signals": {
-    "co_invocation": {
-      "code-reviewer|security-reviewer": 4,
-      "architect-review|test-automator": 3
-    },
-    "skill_cadence": {
-      "python-patterns": {"loads": 12, "sessions": 12}
-    },
-    "file_types": {"py": 87, "md": 31, "tf": 0},
-    "commit_types": {"fix": 8, "feat": 2}
-  },
-  "suggestions": [
-    {
-      "id": "bundle:reviewers-pair",
-      "rationale": "4 co-invocations of code-reviewer + security-reviewer",
-      "evidence_count": 4
-    }
-  ]
-}
-```
-
-## Digest
-
-On `session-end`, the hook calls `format_digest(profile)` and prints
-anything new. Example output:
-
-```
-[behavior-miner] 2 suggestions:
-  - bundle:reviewers-pair  (4 co-invocations)
-    → add to 'review' toolbox:
-       ctx-toolbox add review --post code-reviewer,security-reviewer
-  - promote:python-patterns  (loaded in 12/12 sessions)
-    → promote to 'pre' in your default toolbox
-```
-
-Suggestions are never applied automatically. The user runs the command,
-or accepts the suggestion via `toolbox init --accept <id>`.
-
-## CLI
-
-```bash
-# Rebuild the profile from scratch (scans ~/.claude/history/)
-python -m behavior_miner build
-
-# Show current suggestions
-python -m behavior_miner show
-
-# Print digest (same output as session-end hook)
-python -m behavior_miner digest
-
-# Drop a suggestion (noise reduction)
-python -m behavior_miner dismiss bundle:reviewers-pair
-```
-
-## Privacy
-
-All signal data stays in `~/.claude/`. Nothing is sent over the network.
-The miner never reads file contents — only names, extensions, and commit
-message prefixes.
-
-## Related
-
-- [Intent interview](intent-interview.md) — surfaces miner suggestions
-  during the `toolbox init` flow.
+# Behavior miner
+
+[`src/behavior_miner.py`](https://github.com/stevesolun/ctx/blob/main/src/behavior_miner.py)
+watches your invocation patterns and proposes toolbox tweaks grounded in
+real evidence.
+
+## What it collects
+
+Four signal families, each with `MIN_EVIDENCE = 3` before a suggestion
+can surface:
+
+| Signal | Source | Example suggestion |
+|---|---|---|
+| **Co-invocation** | Pairs of agents invoked in the same session | "You ran `code-reviewer` + `security-reviewer` together 4 times — consider a bundle." |
+| **Skill cadence** | Skill load frequency over time | "`python-patterns` loaded every session — promote to `pre`." |
+| **File-type** | File extensions of work-in-progress | "60% of your diffs touch `.tf` files — consider a Terraform toolbox." |
+| **Commit-type** | Conventional Commit parsing | "8 of your last 10 commits are `fix:` — consider a pre-commit test toolbox." |
+
+## User profile
+
+Signals aggregate into `~/.claude/user-profile.json`:
+
+```jsonc
+{
+  "version": 1,
+  "updated_at": 1713456789,
+  "signals": {
+    "co_invocation": {
+      "code-reviewer|security-reviewer": 4,
+      "architect-review|test-automator": 3
+    },
+    "skill_cadence": {
+      "python-patterns": {"loads": 12, "sessions": 12}
+    },
+    "file_types": {"py": 87, "md": 31, "tf": 0},
+    "commit_types": {"fix": 8, "feat": 2}
+  },
+  "suggestions": [
+    {
+      "id": "bundle:reviewers-pair",
+      "rationale": "4 co-invocations of code-reviewer + security-reviewer",
+      "evidence_count": 4
+    }
+  ]
+}
+```
+
+## Digest
+
+On `session-end`, the hook calls `format_digest(profile)` and prints
+anything new. Example output:
+
+```
+[behavior-miner] 2 suggestions:
+  - bundle:reviewers-pair  (4 co-invocations)
+    → add to 'review' toolbox:
+       ctx-toolbox add review --post code-reviewer,security-reviewer
+  - promote:python-patterns  (loaded in 12/12 sessions)
+    → promote to 'pre' in your default toolbox
+```
+
+Suggestions are never applied automatically. The user runs the command,
+or accepts the suggestion via `toolbox init --accept <id>`.
+
+## CLI
+
+```bash
+# Rebuild the profile from scratch (scans ~/.claude/history/)
+python -m behavior_miner build
+
+# Show current suggestions
+python -m behavior_miner show
+
+# Print digest (same output as session-end hook)
+python -m behavior_miner digest
+
+# Drop a suggestion (noise reduction)
+python -m behavior_miner dismiss bundle:reviewers-pair
+```
+
+## Privacy
+
+All signal data stays in `~/.claude/`. Nothing is sent over the network.
+The miner never reads file contents — only names, extensions, and commit
+message prefixes.
+
+## Related
+
+- [Intent interview](intent-interview.md) — surfaces miner suggestions
+  during the `toolbox init` flow.

docs/toolbox/configuration.md

@@ -1,143 +1,143 @@
-# Configuration
-
-Toolbox config lives in two files:
-
-| Layer | Path | Format | Scope |
-|---|---|---|---|
-| **Global** | `~/.claude/toolboxes.json` | JSON | Every repo on this machine |
-| **Per-repo** | `.toolbox.yaml` (project root) | YAML | This repo only, overrides global |
-
-Per-repo entries shadow global entries with the same name. Fields absent
-from the per-repo file fall back to the global value.
-
-## Schema
-
-```jsonc
-{
-  "version": 1,
-  "toolboxes": {
-    "<name>": {
-      "description": "human-readable purpose",
-
-      // Skills to load before the trigger fires
-      "pre": ["python-patterns", "docs-lookup"],
-
-      // Agents to run after
-      "post": [
-        "code-reviewer",
-        "security-reviewer",
-        "architect-review"
-      ],
-
-      "scope": {
-        // "diff" | "dynamic" | "full"
-        "analysis": "dynamic",
-        // Optional: restrict to these glob projects
-        "projects": ["*"],
-        // Optional: restrict to these file globs
-        "files": ["src/**/*.py"]
-      },
-
-      "budget": {
-        "max_tokens": 60000,
-        "max_seconds": 180
-      },
-
-      "dedup": {
-        // "fresh" = always re-run, "user-configurable" = skip
-        // if same files already reviewed this session
-        "policy": "user-configurable",
-        "window_seconds": 3600
-      },
-
-      "trigger": {
-        "slash": true,
-        "session_start": false,
-        "file_save": false,
-        "pre_commit": true,
-        "session_end": false
-      },
-
-      // If true, HIGH/CRITICAL verdicts block pre-commit
-      "guardrail": true
-    }
-  }
-}
-```
-
-## Field reference
-
-### `pre` and `post`
-
-- `pre` — skills to load before work starts. Loaded into the session's
-  skill manifest, unloaded when the session ends.
-- `post` — agents to invoke after the trigger. Each runs in its own
-  sub-agent context window.
-
-Either list can be empty. A toolbox with only `pre` is a skill preloader;
-one with only `post` is a review council.
-
-### `scope.analysis`
-
-Controls what files the council sees:
-
-| Value | Behavior |
-|---|---|
-| `diff` | Only files with uncommitted changes. Cheapest, fastest. |
-| `dynamic` | Diff + import graph blast radius. Catches downstream regressions. |
-| `full` | Every tracked file. Most thorough; expensive — reserve for security sweeps. |
-
-### `budget`
-
-Enforced by `council_runner`. When the plan would exceed `max_tokens`, the
-runner truncates the file list; when time exceeds `max_seconds`, the
-trigger exits 0 without running remaining agents.
-
-### `dedup`
-
-`fresh` always re-runs. `user-configurable` skips a council run when the
-same file set was reviewed within `window_seconds`. Dedup state lives at
-`~/.claude/toolbox-runs/<plan_hash>.json`.
-
-### `trigger`
-
-At least one trigger must be true. Multiple triggers are allowed — a
-`ship-it` toolbox typically enables `slash`, `pre_commit`, and
-`session_end`.
-
-### `guardrail`
-
-When `true` and the trigger is `pre_commit`, the hook reads
-`<plan_hash>.verdict.json` after the council runs and exits `2` (blocks
-the commit) if level is `HIGH` or `CRITICAL`. See
-[Verdicts & guardrails](verdicts.md).
-
-## Editing tools
-
-```bash
-# List all toolboxes, both layers merged
-ctx-toolbox list
-
-# Show resolved config for one toolbox
-ctx-toolbox show ship-it
-
-# Activate a starter preset
-ctx-toolbox activate ship-it
-
-# Export merged config
-ctx-toolbox export > my-toolboxes.yaml
-
-# Import from file
-ctx-toolbox import my-toolboxes.yaml
-```
-
-## Validation
-
-`toolbox_config.load()` validates on read:
-
-- `version` must equal `1`.
-- Every toolbox needs at least one trigger.
-- `scope.analysis` must be one of `diff`, `dynamic`, `full`.
-- `budget.max_tokens` and `budget.max_seconds` must be positive ints.
-
-Invalid entries raise `ValueError` with the offending key.
+# Configuration
+
+Toolbox config lives in two files:
+
+| Layer | Path | Format | Scope |
+|---|---|---|---|
+| **Global** | `~/.claude/toolboxes.json` | JSON | Every repo on this machine |
+| **Per-repo** | `.toolbox.yaml` (project root) | YAML | This repo only, overrides global |
+
+Per-repo entries shadow global entries with the same name. Fields absent
+from the per-repo file fall back to the global value.
+
+## Schema
+
+```jsonc
+{
+  "version": 1,
+  "toolboxes": {
+    "<name>": {
+      "description": "human-readable purpose",
+
+      // Skills to load before the trigger fires
+      "pre": ["python-patterns", "docs-lookup"],
+
+      // Agents to run after
+      "post": [
+        "code-reviewer",
+        "security-reviewer",
+        "architect-review"
+      ],
+
+      "scope": {
+        // "diff" | "dynamic" | "full"
+        "analysis": "dynamic",
+        // Optional: restrict to these glob projects
+        "projects": ["*"],
+        // Optional: restrict to these file globs
+        "files": ["src/**/*.py"]
+      },
+
+      "budget": {
+        "max_tokens": 60000,
+        "max_seconds": 180
+      },
+
+      "dedup": {
+        // "fresh" = always re-run, "user-configurable" = skip
+        // if same files already reviewed this session
+        "policy": "user-configurable",
+        "window_seconds": 3600
+      },
+
+      "trigger": {
+        "slash": true,
+        "session_start": false,
+        "file_save": false,
+        "pre_commit": true,
+        "session_end": false
+      },
+
+      // If true, HIGH/CRITICAL verdicts block pre-commit
+      "guardrail": true
+    }
+  }
+}
+```
+
+## Field reference
+
+### `pre` and `post`
+
+- `pre` — skills to load before work starts. Loaded into the session's
+  skill manifest, unloaded when the session ends.
+- `post` — agents to invoke after the trigger. Each runs in its own
+  sub-agent context window.
+
+Either list can be empty. A toolbox with only `pre` is a skill preloader;
+one with only `post` is a review council.
+
+### `scope.analysis`
+
+Controls what files the council sees:
+
+| Value | Behavior |
+|---|---|
+| `diff` | Only files with uncommitted changes. Cheapest, fastest. |
+| `dynamic` | Diff + import graph blast radius. Catches downstream regressions. |
+| `full` | Every tracked file. Most thorough; expensive — reserve for security sweeps. |
+
+### `budget`
+
+Enforced by `council_runner`. When the plan would exceed `max_tokens`, the
+runner truncates the file list; when time exceeds `max_seconds`, the
+trigger exits 0 without running remaining agents.
+
+### `dedup`
+
+`fresh` always re-runs. `user-configurable` skips a council run when the
+same file set was reviewed within `window_seconds`. Dedup state lives at
+`~/.claude/toolbox-runs/<plan_hash>.json`.
+
+### `trigger`
+
+At least one trigger must be true. Multiple triggers are allowed — a
+`ship-it` toolbox typically enables `slash`, `pre_commit`, and
+`session_end`.
+
+### `guardrail`
+
+When `true` and the trigger is `pre_commit`, the hook reads
+`<plan_hash>.verdict.json` after the council runs and exits `2` (blocks
+the commit) if level is `HIGH` or `CRITICAL`. See
+[Verdicts & guardrails](verdicts.md).
+
+## Editing tools
+
+```bash
+# List all toolboxes, both layers merged
+ctx-toolbox list
+
+# Show resolved config for one toolbox
+ctx-toolbox show ship-it
+
+# Activate a starter preset
+ctx-toolbox activate ship-it
+
+# Export merged config
+ctx-toolbox export > my-toolboxes.yaml
+
+# Import from file
+ctx-toolbox import my-toolboxes.yaml
+```
+
+## Validation
+
+`toolbox_config.load()` validates on read:
+
+- `version` must equal `1`.
+- Every toolbox needs at least one trigger.
+- `scope.analysis` must be one of `diff`, `dynamic`, `full`.
+- `budget.max_tokens` and `budget.max_seconds` must be positive ints.
+
+Invalid entries raise `ValueError` with the offending key.

docs/toolbox/council-runner.md

@@ -1,81 +1,81 @@
-# Council runner
-
-[`src/council_runner.py`](https://github.com/stevesolun/ctx/blob/main/src/council_runner.py)
-is the planner that turns a toolbox declaration into a concrete `RunPlan`
-the hook system can execute.
-
-## Responsibilities
-
-1. **Resolve the toolbox** — merge global + per-repo config.
-2. **Compute scope** — walk the current diff or full repo, honoring
-   `scope.analysis` and optional `scope.files` globs.
-3. **Graph-blast expansion** — for `dynamic` scope, add every file that
-   imports a changed module (via the knowledge graph edge map).
-4. **Enforce budget** — drop files until the plan fits within
-   `budget.max_tokens` (estimated by line count × heuristic).
-5. **Honor dedup** — skip if the same file set was run within
-   `dedup.window_seconds` and policy is `user-configurable`.
-6. **Persist** — write the plan to
-   `~/.claude/toolbox-runs/<plan_hash>.json` for downstream reads.
-
-## RunPlan
-
-```python
-@dataclass(frozen=True)
-class RunPlan:
-    plan_hash: str
-    toolbox: str
-    agents: tuple[str, ...]
-    files: tuple[str, ...]
-    source: str           # "slash" | "pre-commit" | ...
-    guardrail: bool
-    budget: Budget
-    created_at: float
-```
-
-The `plan_hash` is deterministic (sha256 of `toolbox|sorted(files)|agents`),
-which lets dedup work across triggers without any additional state.
-
-## CLI
-
-```bash
-# Build and persist a plan for the named toolbox
-python -m council_runner build ship-it
-
-# Build without persisting (useful for inspection)
-python -m council_runner build ship-it --dry-run
-
-# Show a previously persisted plan
-python -m council_runner show <plan_hash>
-
-# List recent plans
-python -m council_runner list --limit 10
-```
-
-## Budget estimation
-
-Token estimates are intentionally rough. The runner assumes ~4 tokens per
-line of source, then sorts files by recency (newest first) and greedily
-takes until `max_tokens` is reached. If a single file exceeds the budget,
-the plan is truncated rather than dropped — the council still runs on a
-partial view.
-
-This cheap estimate is fine because the council itself enforces its own
-budgets; `council_runner`'s job is just to stay in the right ballpark.
-
-## Dedup window
-
-Dedup compares the sorted file list, not the plan hash — that way a
-toolbox and its re-run with a newer budget still dedup correctly.
-
-## Graph-blast expansion
-
-For `dynamic` scope, `council_runner` reads the graph edge map produced
-by `scan_repo.py` and walks imports one hop out from each changed file.
-It stops at one hop to keep scope bounded; deep graph walks are reserved
-for explicit `full` mode.
-
-## Related
-
-- [Hooks & triggers](hooks.md) — how a plan gets executed.
-- [Verdicts & guardrails](verdicts.md) — what the council leaves behind.
+# Council runner
+
+[`src/council_runner.py`](https://github.com/stevesolun/ctx/blob/main/src/council_runner.py)
+is the planner that turns a toolbox declaration into a concrete `RunPlan`
+the hook system can execute.
+
+## Responsibilities
+
+1. **Resolve the toolbox** — merge global + per-repo config.
+2. **Compute scope** — walk the current diff or full repo, honoring
+   `scope.analysis` and optional `scope.files` globs.
+3. **Graph-blast expansion** — for `dynamic` scope, add every file that
+   imports a changed module (via the knowledge graph edge map).
+4. **Enforce budget** — drop files until the plan fits within
+   `budget.max_tokens` (estimated by line count × heuristic).
+5. **Honor dedup** — skip if the same file set was run within
+   `dedup.window_seconds` and policy is `user-configurable`.
+6. **Persist** — write the plan to
+   `~/.claude/toolbox-runs/<plan_hash>.json` for downstream reads.
+
+## RunPlan
+
+```python
+@dataclass(frozen=True)
+class RunPlan:
+    plan_hash: str
+    toolbox: str
+    agents: tuple[str, ...]
+    files: tuple[str, ...]
+    source: str           # "slash" | "pre-commit" | ...
+    guardrail: bool
+    budget: Budget
+    created_at: float
+```
+
+The `plan_hash` is deterministic (sha256 of `toolbox|sorted(files)|agents`),
+which lets dedup work across triggers without any additional state.
+
+## CLI
+
+```bash
+# Build and persist a plan for the named toolbox
+python -m council_runner build ship-it
+
+# Build without persisting (useful for inspection)
+python -m council_runner build ship-it --dry-run
+
+# Show a previously persisted plan
+python -m council_runner show <plan_hash>
+
+# List recent plans
+python -m council_runner list --limit 10
+```
+
+## Budget estimation
+
+Token estimates are intentionally rough. The runner assumes ~4 tokens per
+line of source, then sorts files by recency (newest first) and greedily
+takes until `max_tokens` is reached. If a single file exceeds the budget,
+the plan is truncated rather than dropped — the council still runs on a
+partial view.
+
+This cheap estimate is fine because the council itself enforces its own
+budgets; `council_runner`'s job is just to stay in the right ballpark.
+
+## Dedup window
+
+Dedup compares the sorted file list, not the plan hash — that way a
+toolbox and its re-run with a newer budget still dedup correctly.
+
+## Graph-blast expansion
+
+For `dynamic` scope, `council_runner` reads the graph edge map produced
+by `scan_repo.py` and walks imports one hop out from each changed file.
+It stops at one hop to keep scope bounded; deep graph walks are reserved
+for explicit `full` mode.
+
+## Related
+
+- [Hooks & triggers](hooks.md) — how a plan gets executed.
+- [Verdicts & guardrails](verdicts.md) — what the council leaves behind.

docs/toolbox/index.md

@@ -1,75 +1,75 @@
-# Toolbox overview
-
-A **toolbox** is a named bundle of skills and agents that runs at a defined
-moment in your workflow: at session start, on file save, before a commit, at
-session end, or when you invoke its slash command.
-
-Toolboxes let you declare the *council* you want reviewing your work
-without hand-loading skills each session.
-
-## Lifecycle
-
-```mermaid
-flowchart LR
-  A[Declare toolbox] --> B[Trigger fires]
-  B --> C[Council runner<br/>builds plan]
-  C --> D[Agents run<br/>scoped to plan.files]
-  D --> E[Findings recorded<br/>as Verdict]
-  E -->|HIGH / CRITICAL| F[Guardrail blocks<br/>pre-commit]
-  E -->|LOW / MEDIUM| G[Logged,<br/>session continues]
-```
-
-Each arrow is a concrete module:
-
-- **Declare**: [`toolbox_config.py`](https://github.com/stevesolun/ctx/blob/main/src/toolbox_config.py)
-  loads `~/.claude/toolboxes.json` and merges per-repo `.toolbox.yaml` on top.
-- **Trigger**: [`toolbox_hooks.py`](https://github.com/stevesolun/ctx/blob/main/src/toolbox_hooks.py)
-  listens for `session-start`, `file-save`, `pre-commit`, `session-end`, and
-  the `/toolbox run` slash command.
-- **Plan**: [`council_runner.py`](https://github.com/stevesolun/ctx/blob/main/src/council_runner.py)
-  assembles a `RunPlan` honoring scope, dedup, and graph-blast expansion.
-- **Verdict**: [`toolbox_verdict.py`](https://github.com/stevesolun/ctx/blob/main/src/toolbox_verdict.py)
-  merges findings by id and escalates level to max(findings).
-
-## Minimal declaration
-
-```yaml
-# .toolbox.yaml (per-repo)
-version: 1
-toolboxes:
-  review:
-    description: "Post-feature code review"
-    post:
-      - code-reviewer
-      - security-reviewer
-    scope:
-      analysis: diff
-    trigger:
-      slash: true
-      pre_commit: true
-    guardrail: true
-```
-
+# Toolbox overview
+
+A **toolbox** is a named bundle of skills and agents that runs at a defined
+moment in your workflow: at session start, on file save, before a commit, at
+session end, or when you invoke its slash command.
+
+Toolboxes let you declare the *council* you want reviewing your work
+without hand-loading skills each session.
+
+## Lifecycle
+
+```mermaid
+flowchart LR
+  A[Declare toolbox] --> B[Trigger fires]
+  B --> C[Council runner<br/>builds plan]
+  C --> D[Agents run<br/>scoped to plan.files]
+  D --> E[Findings recorded<br/>as Verdict]
+  E -->|HIGH / CRITICAL| F[Guardrail blocks<br/>pre-commit]
+  E -->|LOW / MEDIUM| G[Logged,<br/>session continues]
+```
+
+Each arrow is a concrete module:
+
+- **Declare**: [`toolbox_config.py`](https://github.com/stevesolun/ctx/blob/main/src/toolbox_config.py)
+  loads `~/.claude/toolboxes.json` and merges per-repo `.toolbox.yaml` on top.
+- **Trigger**: [`toolbox_hooks.py`](https://github.com/stevesolun/ctx/blob/main/src/toolbox_hooks.py)
+  listens for `session-start`, `file-save`, `pre-commit`, `session-end`, and
+  the `/toolbox run` slash command.
+- **Plan**: [`council_runner.py`](https://github.com/stevesolun/ctx/blob/main/src/council_runner.py)
+  assembles a `RunPlan` honoring scope, dedup, and graph-blast expansion.
+- **Verdict**: [`toolbox_verdict.py`](https://github.com/stevesolun/ctx/blob/main/src/toolbox_verdict.py)
+  merges findings by id and escalates level to max(findings).
+
+## Minimal declaration
+
+```yaml
+# .toolbox.yaml (per-repo)
+version: 1
+toolboxes:
+  review:
+    description: "Post-feature code review"
+    post:
+      - code-reviewer
+      - security-reviewer
+    scope:
+      analysis: diff
+    trigger:
+      slash: true
+      pre_commit: true
+    guardrail: true
+```
+
 Run it manually:
 
 ```bash
 ctx-toolbox run --event pre-commit
 ```
-
-Or let the `pre-commit` hook fire it automatically — see
-[Hooks & triggers](hooks.md).
-
-## Scope modes
-
-| Mode | What gets reviewed | Best for |
-|---|---|---|
-| `diff` | Files in the current uncommitted diff | Pre-commit, real-time review |
-| `dynamic` | Diff + graph blast radius (imports of modified files) | Refactor safety |
-| `full` | Entire repo | Security sweeps, docs audits |
-
-## Related
-
-- [Configuration schema](configuration.md) — full field reference.
-- [Starter toolboxes](starters.md) — 5 shipping presets.
-- [Intent interview](intent-interview.md) — `toolbox init` walkthrough.
-- [Verdicts & guardrails](verdicts.md) — how blocking works.
+
+Or let the `pre-commit` hook fire it automatically — see
+[Hooks & triggers](hooks.md).
+
+## Scope modes
+
+| Mode | What gets reviewed | Best for |
+|---|---|---|
+| `diff` | Files in the current uncommitted diff | Pre-commit, real-time review |
+| `dynamic` | Diff + graph blast radius (imports of modified files) | Refactor safety |
+| `full` | Entire repo | Security sweeps, docs audits |
+
+## Related
+
+- [Configuration schema](configuration.md) — full field reference.
+- [Starter toolboxes](starters.md) — 5 shipping presets.
+- [Intent interview](intent-interview.md) — `toolbox init` walkthrough.
+- [Verdicts & guardrails](verdicts.md) — how blocking works.

docs/toolbox/intent-interview.md

@@ -1,94 +1,94 @@
-# Intent interview
-
-[`src/intent_interview.py`](https://github.com/stevesolun/ctx/blob/main/src/intent_interview.py)
-bootstraps your toolbox set via a short, skippable interview.
-
-The slash command `/toolbox init` is a thin wrapper; see
-[`.claude/commands/toolbox-init.md`](https://github.com/stevesolun/ctx/blob/main/.claude/commands/toolbox-init.md).
-
-## Flow
-
-1. **Detect repo state** — is this a git repo? Any commits? What languages?
-2. **Load behavior profile** — read `~/.claude/user-profile.json` for any
-   mined suggestions.
-3. **Ask up to three questions**:
-   - Which starter toolboxes to activate.
-   - Which miner suggestions to accept (if any).
-   - Default analysis mode for new toolboxes.
-4. **Persist** — write chosen toolboxes to `~/.claude/toolboxes.json`
-   (only when `--apply` is passed).
-
-Any prompt can be skipped with the word `skip`.
-
-## Repo state detection
-
-```python
-@dataclass(frozen=True)
-class RepoState:
-    is_git: bool
-    commit_count: int
-    languages: dict[str, int]     # extension → file count
-    markers: dict[str, str]       # marker file → language
-    has_toolbox_config: bool
-
-    @property
-    def is_blank(self) -> bool:
-        # True when the repo has effectively nothing to analyze yet.
-        return not self.is_git or self.commit_count == 0 or (
-            not self.languages and not self.markers
-        )
-```
-
-Language scoring uses both extensions (`.py`, `.ts`, …) and marker files
-(`pyproject.toml`, `Cargo.toml`, `Dockerfile`, …). Marker files bump the
-score by 5 to reflect that they declare intent more strongly than a
-stray extension match.
-
-## Usage
-
-```bash
-# Default: interactive, dry-run (no write)
-python -m intent_interview init
-
-# Detect state only
-python -m intent_interview detect
-
-# Preset flows (no prompts)
-python -m intent_interview init --preset blank --apply
-python -m intent_interview init --preset existing --apply
-python -m intent_interview init --preset docs-heavy --apply
-python -m intent_interview init --preset security-first --apply
-
-# Fully structured (CI / scripted setup)
-python -m intent_interview init \
-  --non-interactive \
-  --starters ship-it,security-sweep \
-  --suggestions 1,2 \
-  --analysis dynamic \
-  --apply
-```
-
-## Presets
-
-| Preset | Starters | Default scope |
-|---|---|---|
-| `blank` | ship-it, security-sweep, fresh-repo-init | dynamic |
-| `existing` | ship-it, refactor-safety | dynamic |
-| `docs-heavy` | docs-review | diff |
-| `security-first` | security-sweep | full |
-
-## Skip semantics
-
-- Typing `skip` at any prompt short-circuits the whole interview: no
-  starters activated, no suggestions accepted, analysis mode unchanged.
-- `--skip` on the CLI is the non-interactive equivalent.
-
-## Exit codes
-
-- `0` — success; JSON payload printed on stdout.
-- non-zero — unrecoverable error (unknown preset, malformed args).
-
-## Related
-
-- [Starter toolboxes](starters.md) — the five bundles the interview can activate.
-- [Behavior miner](behavior-miner.md) — source of the suggestions the interview offers.
+# Intent interview
+
+[`src/intent_interview.py`](https://github.com/stevesolun/ctx/blob/main/src/intent_interview.py)
+bootstraps your toolbox set via a short, skippable interview.
+
+The slash command `/toolbox init` is a thin wrapper; see
+[`.claude/commands/toolbox-init.md`](https://github.com/stevesolun/ctx/blob/main/.claude/commands/toolbox-init.md).
+
+## Flow
+
+1. **Detect repo state** — is this a git repo? Any commits? What languages?
+2. **Load behavior profile** — read `~/.claude/user-profile.json` for any
+   mined suggestions.
+3. **Ask up to three questions**:
+   - Which starter toolboxes to activate.
+   - Which miner suggestions to accept (if any).
+   - Default analysis mode for new toolboxes.
+4. **Persist** — write chosen toolboxes to `~/.claude/toolboxes.json`
+   (only when `--apply` is passed).
+
+Any prompt can be skipped with the word `skip`.
+
+## Repo state detection
+
+```python
+@dataclass(frozen=True)
+class RepoState:
+    is_git: bool
+    commit_count: int
+    languages: dict[str, int]     # extension → file count
+    markers: dict[str, str]       # marker file → language
+    has_toolbox_config: bool
+
+    @property
+    def is_blank(self) -> bool:
+        # True when the repo has effectively nothing to analyze yet.
+        return not self.is_git or self.commit_count == 0 or (
+            not self.languages and not self.markers
+        )
+```
+
+Language scoring uses both extensions (`.py`, `.ts`, …) and marker files
+(`pyproject.toml`, `Cargo.toml`, `Dockerfile`, …). Marker files bump the
+score by 5 to reflect that they declare intent more strongly than a
+stray extension match.
+
+## Usage
+
+```bash
+# Default: interactive, dry-run (no write)
+python -m intent_interview init
+
+# Detect state only
+python -m intent_interview detect
+
+# Preset flows (no prompts)
+python -m intent_interview init --preset blank --apply
+python -m intent_interview init --preset existing --apply
+python -m intent_interview init --preset docs-heavy --apply
+python -m intent_interview init --preset security-first --apply
+
+# Fully structured (CI / scripted setup)
+python -m intent_interview init \
+  --non-interactive \
+  --starters ship-it,security-sweep \
+  --suggestions 1,2 \
+  --analysis dynamic \
+  --apply
+```
+
+## Presets
+
+| Preset | Starters | Default scope |
+|---|---|---|
+| `blank` | ship-it, security-sweep, fresh-repo-init | dynamic |
+| `existing` | ship-it, refactor-safety | dynamic |
+| `docs-heavy` | docs-review | diff |
+| `security-first` | security-sweep | full |
+
+## Skip semantics
+
+- Typing `skip` at any prompt short-circuits the whole interview: no
+  starters activated, no suggestions accepted, analysis mode unchanged.
+- `--skip` on the CLI is the non-interactive equivalent.
+
+## Exit codes
+
+- `0` — success; JSON payload printed on stdout.
+- non-zero — unrecoverable error (unknown preset, malformed args).
+
+## Related
+
+- [Starter toolboxes](starters.md) — the five bundles the interview can activate.
+- [Behavior miner](behavior-miner.md) — source of the suggestions the interview offers.

docs/toolbox/starters.md

@@ -1,96 +1,96 @@
-# Starter toolboxes
-
-Five presets ship in `docs/toolbox/templates/`. `toolbox init` activates
-them into `~/.claude/toolboxes.json`; you can then override any field per-repo
-in `.toolbox.yaml`.
-
-## ship-it
-
-> **Professional council of 7 experts for end-of-feature review.**
-
-Runs `code-reviewer`, `security-reviewer`, `architect-review`,
-`test-automator`, `performance-engineer`, `accessibility-tester`, and
-`docs-lookup` against the diff + graph blast radius.
-
-- **Triggers**: slash, pre-commit, session-end.
-- **Scope**: `dynamic` — diff plus imports of changed modules.
-- **Budget**: 200 k tokens / 420 seconds.
-- **Guardrail**: on (HIGH/CRITICAL blocks pre-commit).
-
-Best for: shipping a feature branch. The council covers correctness,
-security, architecture, testing, performance, accessibility, and docs in
-one pass.
-
-## security-sweep
-
-> **Full-repo security audit with blocking guardrail on HIGH findings.**
-
-Runs `security-reviewer`, `backend-security-coder`, `frontend-security-coder`,
-and `compliance-auditor` against the entire repo.
-
-- **Triggers**: slash, pre-commit.
-- **Scope**: `full` — every tracked file.
-- **Budget**: 300 k tokens / 600 seconds.
-- **Guardrail**: on.
-
-Best for: periodic audits, pre-release sweeps, compliance checkpoints.
-Expensive; not a per-commit hook.
-
-## refactor-safety
-
-> **Graph-informed refactor review with regression and dead-code checks.**
-
-Runs `architect-review`, `refactor-cleaner`, `test-automator`, and
-`code-reviewer` against diff + graph blast.
-
-- **Triggers**: slash, pre-commit.
-- **Scope**: `dynamic`.
-- **Budget**: 120 k tokens / 300 seconds.
-- **Guardrail**: off — flags issues without blocking.
-
-Best for: mid-refactor checkpoints. Catches orphaned code, downstream
-breakage, missing test updates.
-
-## docs-review
-
-> **Documentation pass: accuracy, completeness, clarity, and API parity.**
-
-Runs `docs-lookup`, `technical-writer`, and `code-reviewer` against docs
-diffs.
-
-- **Triggers**: slash, pre-commit.
-- **Scope**: `diff`.
-- **Budget**: 60 k tokens / 180 seconds.
-- **Guardrail**: off.
-
-Best for: docs-heavy branches and README updates.
-
-## fresh-repo-init
-
-> **New-repo bootstrap: run the intent interview, scaffold plan, pick initial toolbox.**
-
-Invokes `intent_interview` in interactive mode, then activates whichever
-starters the user selects.
-
-- **Triggers**: slash only.
-- **Scope**: `full` (fresh scan of a new repo).
-- **Budget**: small — this bundle is just an orchestrator.
-
-Best for: `git init` followed by `toolbox init`.
-
-## Activation
-
-```bash
-# Pick starters interactively
-python -m intent_interview init
-
-# Non-interactive preset
-python -m intent_interview init --preset existing --apply
-python -m intent_interview init --preset docs-heavy --apply
-python -m intent_interview init --preset security-first --apply
-
-# Activate a specific starter directly
-ctx-toolbox activate ship-it
-```
-
-See [Intent interview](intent-interview.md) for the full flow.
+# Starter toolboxes
+
+Five presets ship in `docs/toolbox/templates/`. `toolbox init` activates
+them into `~/.claude/toolboxes.json`; you can then override any field per-repo
+in `.toolbox.yaml`.
+
+## ship-it
+
+> **Professional council of 7 experts for end-of-feature review.**
+
+Runs `code-reviewer`, `security-reviewer`, `architect-review`,
+`test-automator`, `performance-engineer`, `accessibility-tester`, and
+`docs-lookup` against the diff + graph blast radius.
+
+- **Triggers**: slash, pre-commit, session-end.
+- **Scope**: `dynamic` — diff plus imports of changed modules.
+- **Budget**: 200 k tokens / 420 seconds.
+- **Guardrail**: on (HIGH/CRITICAL blocks pre-commit).
+
+Best for: shipping a feature branch. The council covers correctness,
+security, architecture, testing, performance, accessibility, and docs in
+one pass.
+
+## security-sweep
+
+> **Full-repo security audit with blocking guardrail on HIGH findings.**
+
+Runs `security-reviewer`, `backend-security-coder`, `frontend-security-coder`,
+and `compliance-auditor` against the entire repo.
+
+- **Triggers**: slash, pre-commit.
+- **Scope**: `full` — every tracked file.
+- **Budget**: 300 k tokens / 600 seconds.
+- **Guardrail**: on.
+
+Best for: periodic audits, pre-release sweeps, compliance checkpoints.
+Expensive; not a per-commit hook.
+
+## refactor-safety
+
+> **Graph-informed refactor review with regression and dead-code checks.**
+
+Runs `architect-review`, `refactor-cleaner`, `test-automator`, and
+`code-reviewer` against diff + graph blast.
+
+- **Triggers**: slash, pre-commit.
+- **Scope**: `dynamic`.
+- **Budget**: 120 k tokens / 300 seconds.
+- **Guardrail**: off — flags issues without blocking.
+
+Best for: mid-refactor checkpoints. Catches orphaned code, downstream
+breakage, missing test updates.
+
+## docs-review
+
+> **Documentation pass: accuracy, completeness, clarity, and API parity.**
+
+Runs `docs-lookup`, `technical-writer`, and `code-reviewer` against docs
+diffs.
+
+- **Triggers**: slash, pre-commit.
+- **Scope**: `diff`.
+- **Budget**: 60 k tokens / 180 seconds.
+- **Guardrail**: off.
+
+Best for: docs-heavy branches and README updates.
+
+## fresh-repo-init
+
+> **New-repo bootstrap: run the intent interview, scaffold plan, pick initial toolbox.**
+
+Invokes `intent_interview` in interactive mode, then activates whichever
+starters the user selects.
+
+- **Triggers**: slash only.
+- **Scope**: `full` (fresh scan of a new repo).
+- **Budget**: small — this bundle is just an orchestrator.
+
+Best for: `git init` followed by `toolbox init`.
+
+## Activation
+
+```bash
+# Pick starters interactively
+python -m intent_interview init
+
+# Non-interactive preset
+python -m intent_interview init --preset existing --apply
+python -m intent_interview init --preset docs-heavy --apply
+python -m intent_interview init --preset security-first --apply
+
+# Activate a specific starter directly
+ctx-toolbox activate ship-it
+```
+
+See [Intent interview](intent-interview.md) for the full flow.

docs/toolbox/verdicts.md

@@ -1,142 +1,142 @@
-# Verdicts & guardrails
-
-[`src/toolbox_verdict.py`](https://github.com/stevesolun/ctx/blob/main/src/toolbox_verdict.py)
-owns the council's finding ledger. A `RunPlan` says *what should run*; a
-`Verdict` says *what was found* and, if the level escalates high enough,
-blocks `git commit`.
-
-## Data model
-
-```python
-@dataclass(frozen=True)
-class Evidence:
-    file: str
-    line: int | None = None
-    note: str = ""
-
-@dataclass(frozen=True)
-class Finding:
-    id: str                    # stable, hash(level|agent|title)
-    level: str                 # "LOW" | "MEDIUM" | "HIGH" | "CRITICAL"
-    title: str
-    agent: str = ""
-    evidence: tuple[Evidence, ...] = ()
-    rationale: str = ""
-    created_at: float = 0.0
-
-@dataclass(frozen=True)
-class Verdict:
-    plan_hash: str
-    level: str                 # max(findings)
-    summary: str
-    findings: tuple[Finding, ...]
-    created_at: float
-    updated_at: float
-```
-
-Level escalation is always `max(findings)`. Clearing a finding
-re-escalates from whatever remains.
-
-## Storage
-
-```
-~/.claude/toolbox-runs/
-  abc123.json                # the RunPlan
-  abc123.verdict.json        # the Verdict (sibling)
-```
-
-Same directory, same hash. A single `history` sweep covers both.
-
-## Merge-by-id
-
-Agents can refine a previous finding by recording with the same id.
-The new record replaces the old one (rationale, evidence, level all
-update). This is how a `security-reviewer` can start with a MEDIUM
-finding, then bump it to CRITICAL after deeper analysis — without
-leaving duplicate entries.
-
-The default id is `sha256(level|agent|title)[:12]`, so the same
-agent recording the same titled issue naturally dedups. Pass
-`--id custom-value` if you need a different stable key.
-
-## Blocking
-
-`toolbox_hooks` reads `<plan>.verdict.json` after a `pre-commit`
-council runs. If `level in {"HIGH", "CRITICAL"}` and the toolbox has
-`guardrail: true`, it returns exit `2`, which stops the commit.
-
-`LOW` and `MEDIUM` findings are logged but never block.
-
-## CLI
-
-```bash
-# Record a finding
-python -m toolbox_verdict record \
-  --plan-hash abc123 \
-  --level HIGH \
-  --title "SQL injection in users.py" \
-  --agent security-reviewer \
-  --evidence src/users.py:42:unescaped input \
-  --rationale "req.form values flow into raw SQL"
-
-# Show the verdict
-python -m toolbox_verdict show --plan-hash abc123
-
-# JSON payload for piping
-python -m toolbox_verdict show --plan-hash abc123 --json
-
-# Recent verdicts (retrospective)
-python -m toolbox_verdict retro --limit 10
-
-# Only HIGH/CRITICAL
-python -m toolbox_verdict retro --min-level HIGH
-
-# Pretty-print the evidence chain
-python -m toolbox_verdict explain --plan-hash abc123
-
-# Remove a single finding
-python -m toolbox_verdict clear --plan-hash abc123 --id <id>
-```
-
-## Evidence parsing
-
-`parse_evidence()` accepts three forms, parsed right-to-left so Windows
-drive-letter paths don't trip the delimiter:
-
-- `src/foo.py` → `Evidence(file="src/foo.py")`
-- `src/foo.py:42` → `Evidence(file="src/foo.py", line=42)`
-- `src/foo.py:42:race on counter` → `Evidence(file=..., line=42, note=...)`
-- `C:/Users/me/foo.py:17` → `Evidence(file="C:/Users/me/foo.py", line=17)`
-
-Empty specs yield an `Evidence` with an empty file and are filtered out
-at `build_finding()` time.
-
-## Explain output
-
-```
-[verdict] plan=abc123  level=HIGH  2 finding(s): 1 high, 1 low
-  - [HIGH] SQL injection in users.py  (agent: security-reviewer)
-      why: req.form values flow into raw SQL
-      evidence: src/users.py:42 — unescaped input
-  - [LOW] style: trailing whitespace  (agent: code-reviewer)
-      evidence: src/users.py:57
-```
-
-Findings render in severity-desc order so the blocking issue always
-appears first.
-
-## Retrospective
-
-`recent_verdicts()` returns verdicts sorted by `updated_at` desc:
-
-```
-[retro] 3 recent verdict(s):
-  - plan-crit   CRITICAL  BLOCK  1 finding(s): 1 critical
-  - plan-hi     HIGH      BLOCK  2 finding(s): 1 high, 1 low
-  - plan-ok     LOW       ok     1 finding(s): 1 low
-```
-
-## Related
-
-- [Hooks & triggers](hooks.md) — where the `2` exit blocks the commit.
-- [Council runner](council-runner.md) — where the plan hash comes from.
+# Verdicts & guardrails
+
+[`src/toolbox_verdict.py`](https://github.com/stevesolun/ctx/blob/main/src/toolbox_verdict.py)
+owns the council's finding ledger. A `RunPlan` says *what should run*; a
+`Verdict` says *what was found* and, if the level escalates high enough,
+blocks `git commit`.
+
+## Data model
+
+```python
+@dataclass(frozen=True)
+class Evidence:
+    file: str
+    line: int | None = None
+    note: str = ""
+
+@dataclass(frozen=True)
+class Finding:
+    id: str                    # stable, hash(level|agent|title)
+    level: str                 # "LOW" | "MEDIUM" | "HIGH" | "CRITICAL"
+    title: str
+    agent: str = ""
+    evidence: tuple[Evidence, ...] = ()
+    rationale: str = ""
+    created_at: float = 0.0
+
+@dataclass(frozen=True)
+class Verdict:
+    plan_hash: str
+    level: str                 # max(findings)
+    summary: str
+    findings: tuple[Finding, ...]
+    created_at: float
+    updated_at: float
+```
+
+Level escalation is always `max(findings)`. Clearing a finding
+re-escalates from whatever remains.
+
+## Storage
+
+```
+~/.claude/toolbox-runs/
+  abc123.json                # the RunPlan
+  abc123.verdict.json        # the Verdict (sibling)
+```
+
+Same directory, same hash. A single `history` sweep covers both.
+
+## Merge-by-id
+
+Agents can refine a previous finding by recording with the same id.
+The new record replaces the old one (rationale, evidence, level all
+update). This is how a `security-reviewer` can start with a MEDIUM
+finding, then bump it to CRITICAL after deeper analysis — without
+leaving duplicate entries.
+
+The default id is `sha256(level|agent|title)[:12]`, so the same
+agent recording the same titled issue naturally dedups. Pass
+`--id custom-value` if you need a different stable key.
+
+## Blocking
+
+`toolbox_hooks` reads `<plan>.verdict.json` after a `pre-commit`
+council runs. If `level in {"HIGH", "CRITICAL"}` and the toolbox has
+`guardrail: true`, it returns exit `2`, which stops the commit.
+
+`LOW` and `MEDIUM` findings are logged but never block.
+
+## CLI
+
+```bash
+# Record a finding
+python -m toolbox_verdict record \
+  --plan-hash abc123 \
+  --level HIGH \
+  --title "SQL injection in users.py" \
+  --agent security-reviewer \
+  --evidence src/users.py:42:unescaped input \
+  --rationale "req.form values flow into raw SQL"
+
+# Show the verdict
+python -m toolbox_verdict show --plan-hash abc123
+
+# JSON payload for piping
+python -m toolbox_verdict show --plan-hash abc123 --json
+
+# Recent verdicts (retrospective)
+python -m toolbox_verdict retro --limit 10
+
+# Only HIGH/CRITICAL
+python -m toolbox_verdict retro --min-level HIGH
+
+# Pretty-print the evidence chain
+python -m toolbox_verdict explain --plan-hash abc123
+
+# Remove a single finding
+python -m toolbox_verdict clear --plan-hash abc123 --id <id>
+```
+
+## Evidence parsing
+
+`parse_evidence()` accepts three forms, parsed right-to-left so Windows
+drive-letter paths don't trip the delimiter:
+
+- `src/foo.py` → `Evidence(file="src/foo.py")`
+- `src/foo.py:42` → `Evidence(file="src/foo.py", line=42)`
+- `src/foo.py:42:race on counter` → `Evidence(file=..., line=42, note=...)`
+- `C:/Users/me/foo.py:17` → `Evidence(file="C:/Users/me/foo.py", line=17)`
+
+Empty specs yield an `Evidence` with an empty file and are filtered out
+at `build_finding()` time.
+
+## Explain output
+
+```
+[verdict] plan=abc123  level=HIGH  2 finding(s): 1 high, 1 low
+  - [HIGH] SQL injection in users.py  (agent: security-reviewer)
+      why: req.form values flow into raw SQL
+      evidence: src/users.py:42 — unescaped input
+  - [LOW] style: trailing whitespace  (agent: code-reviewer)
+      evidence: src/users.py:57
+```
+
+Findings render in severity-desc order so the blocking issue always
+appears first.
+
+## Retrospective
+
+`recent_verdicts()` returns verdicts sorted by `updated_at` desc:
+
+```
+[retro] 3 recent verdict(s):
+  - plan-crit   CRITICAL  BLOCK  1 finding(s): 1 critical
+  - plan-hi     HIGH      BLOCK  2 finding(s): 1 high, 1 low
+  - plan-ok     LOW       ok     1 finding(s): 1 low
+```
+
+## Related
+
+- [Hooks & triggers](hooks.md) — where the `2` exit blocks the commit.
+- [Council runner](council-runner.md) — where the plan hash comes from.

graph/README.md

@@ -1,274 +1,274 @@
-# Knowledge Graph Artifacts
-
-This directory ships the pre-built ctx LLM-wiki and knowledge graph.
-
-Current snapshot:
-
-- **102,928 graph nodes**
-- **2,913,960 graph edges**
-- **52 Louvain communities**
-- **91,464 skill entity pages**; **89,465** have hydrated catalog bodies
-- **467 agent pages**
-- **10,790 MCP server pages**
-- **207 harness pages**
-- **89,465 hydrated `SKILL.md` bodies**
-- **28,612 long skill bodies converted through the micro-skill gate**
-
-The runtime recommendation paths use this graph in two ways:
-
-- Development recommendations return skills, agents, and MCP servers only.
-- Custom/API/local model onboarding recommends harnesses using the higher harness fit floor in `src/config.json`.
-
-## Files
-
-| File | Contents |
-|---|---|
-| `wiki-graph-runtime.tar.gz` | Fast install artifact used by default `ctx-init --graph`: `graphify-out/*`, the skill index, 207 harness pages, wiki index files, and Obsidian metadata needed for recommendations and harness dry-runs without expanding every entity page |
-| `wiki-graph.tar.gz` | Full LLM-wiki: entity pages, converted skill bodies, mirrored agent bodies, concept pages, `graphify-out/graph.json`, `graph-delta.json`, export manifest, communities, skill indexes, and Obsidian metadata |
-| Skill catalog gzip | Compressed skill index for the 89,465 body-backed skill entries shipped in the wiki |
-| `communities.json` | Current Louvain community export |
-| `entity-overlays.jsonl` | Release overlay for first-class entities added after the base graph export; installed beside `graphify-out/graph.json` by `ctx-init --graph` |
-| `graphify-out/dashboard-neighborhoods.sqlite3` inside both tarballs | Compact top-neighbor index used by `ctx-monitor` so `/api/graph/<slug>.json` does not cold-parse the 818 MiB graph JSON |
-| `viz-overview.html` | Plotly overview of the graph |
-| `viz-python.html` | Python-focused graph view |
-| `viz-security.html` | Security-focused graph view |
-| `viz-ai-agents.html` | AI-agent-focused graph view |
-| `sample-top60.html` | Interactive top-degree sample |
-
-Preview HTML files are generated from the shipped `graphify-out/graph.json`
-and embed the graph export ID in `<meta name="ctx-graph-export-id">`. Static
-PNG snapshots are intentionally not shipped because they can drift from the
-current tarball without an executable freshness check.
-
-## Runtime vs Full Wiki
-
-`ctx-init --graph` installs `wiki-graph-runtime.tar.gz` by default. That is the
-right path for recommendations and first-time installs because it avoids
-expanding hundreds of thousands of markdown files while still shipping the
-harness pages needed by `ctx-harness-install --dry-run`. Use
-`ctx-init --graph --graph-install-mode full` or manual full extraction when you
-want local wiki browsing, Obsidian, or the converted skill body tree.
-
-## What Is Inside `wiki-graph.tar.gz`
-
-- `entities/skills/` - all skill entity pages
-- `entities/agents/` - agent entity pages
-- `entities/mcp-servers/<shard>/` - sharded MCP server entity pages
-- `entities/harnesses/` - harness entity pages
-- `converted/` - installable skill bodies
-- `converted-agents/` - mirrored agent bodies
-- `concepts/` - community concept pages
-- `external-catalogs/` - machine-readable skill index, summary, and coverage metadata
-- `graphify-out/graph.json` - NetworkX node-link graph
-- `graphify-out/graph-delta.json` - delta export for the latest graph generation
-- `graphify-out/graph-export-manifest.json` - export manifest tying graph, delta, communities, and report to one generation
-- `graphify-out/communities.json` - community export
-- `SCHEMA.md`, `index.md`, `log.md`, `catalog.md` - wiki contract and indexes
-- `.obsidian/` - vault metadata for local graph browsing
-
-`SKILL.md.original` backups, transient `.lock` files, and `.ctx/` queue state
-are not shipped. Local micro-skill conversion may keep `.original` files for
-traceability, but the packaged tarball excludes them so users do not ingest raw
-long bodies after conversion.
-
-## Extract
-
-Default runtime install:
-
-```bash
-ctx-init --graph
-```
-
-Full wiki extraction:
-
-```bash
-mkdir -p ~/.claude/skill-wiki
-tar xzf graph/wiki-graph.tar.gz -C ~/.claude/skill-wiki/
-```
-
-On Windows PowerShell, use the built-in `tar.exe` without `--force-local`:
-
-```powershell
-New-Item -ItemType Directory -Force "$env:USERPROFILE\.claude\skill-wiki"
-tar -xzf graph\wiki-graph.tar.gz -C "$env:USERPROFILE\.claude\skill-wiki"
-```
-
-With Git Bash or MSYS tar, use `--force-local` only when the `-C` target is a
-drive-letter path:
-
-```bash
-tar --force-local xzf graph/wiki-graph.tar.gz -C C:/Users/<you>/.claude/skill-wiki/
-```
-
-## Validate
-
-```bash
-python src/validate_graph_artifacts.py --deep
-python src/update_repo_stats.py --check
-```
-
-For release-count validation, pin the exact snapshot numbers:
-
-```bash
-python src/validate_graph_artifacts.py --deep \
-  --expected-nodes 102928 \
-  --expected-edges 2913960 \
-  --expected-semantic-edges 1683193 \
-  --expected-harness-nodes 207 \
-  --expected-skill-pages 91464 \
-  --expected-agent-pages 467 \
-  --expected-mcp-pages 10790 \
-  --expected-harness-pages 207
-```
-
-Manual sanity checks:
-
-```bash
-tar -tzf graph/wiki-graph.tar.gz | grep 'graphify-out/graph.json'
-tar -tzf graph/wiki-graph.tar.gz | grep 'external-catalogs/.*/catalog.json'
-tar -tzf graph/wiki-graph.tar.gz | grep 'SKILL.md.original' && exit 1 || true
-tar -tzf graph/wiki-graph.tar.gz | grep '\.lock$' && exit 1 || true
-tar -tzf graph/wiki-graph.tar.gz | grep '^\./\.ctx/' && exit 1 || true
-```
-
-Windows PowerShell equivalent for the exclusion checks:
-
-```powershell
-tar -tzf graph/wiki-graph.tar.gz | Select-String 'SKILL.md.original'
-tar -tzf graph/wiki-graph.tar.gz | Select-String '\.lock$'
-tar -tzf graph/wiki-graph.tar.gz | Select-String '^\./\.ctx/'
-```
-
-The PowerShell commands should print nothing.
-
-## Rebuild
-
-After adding or updating skills, agents, MCP servers, or harnesses:
-
-```bash
-ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1
-ctx-scan-repo --repo . --recommend
-```
-
-The worker path is the fast local update path. It validates the queued entity
-page, updates the wiki index, and attempts incremental ANN attach into
-`graphify-out/entity-overlays.jsonl` when the semantic vector index exists. It
-also queues the normal incremental graph export job, so a full rebuild remains
-the reconciliation path for release artifacts.
-
-If the worker reports that incremental attach was skipped because no vector
-index exists, build the exact portable index:
-
-```bash
-ctx-wiki-graphify \
-  --wiki-dir ~/.claude/skill-wiki \
-  --incremental \
-  --graph-only \
-  --semantic-vector-index numpy-flat
-```
-
-Then drain pending queue work again:
-
-```bash
-ctx-wiki-worker --wiki ~/.claude/skill-wiki
-```
-
-Before promoting an ANN backend or changed thresholds, run the shadow gate:
-
-```bash
-ctx-incremental-shadow \
-  --index-dir ~/.claude/skill-wiki/.embedding-cache/graph/vector-index \
-  --graph ~/.claude/skill-wiki/graphify-out/graph.json \
-  --sample-size 100 \
-  --min-overlap 0.85
-```
-
-It reports precision/recall, top-k agreement, score deltas, and bad examples;
-the release gate fails when recall at the largest requested top-k is below the
-overlap floor.
-
-For release artifact rebuilds:
-
-```bash
-python scripts/graph_artifact_guard.py park
-ctx-wiki-graphify
-python src/validate_graph_artifacts.py --deep
-python src/update_repo_stats.py --check
-```
-
-`park` sets Git's local `skip-worktree` bit for the heavyweight generated
-archives: `graph/wiki-graph.tar.gz`, `graph/wiki-graph-runtime.tar.gz`, and
-the compressed skill index. Keep them parked while graph/wiki generation,
-validation, dashboard smoke, and stats checks are still in progress. This
-prevents background Git integrations from repeatedly staging hundreds of
-megabytes through the Git LFS clean filter. When the release candidate is final,
-unpark and stage the artifacts exactly once:
-
-```bash
-python scripts/graph_artifact_guard.py unpark
-git add graph/wiki-graph.tar.gz graph/wiki-graph-runtime.tar.gz \
-  graph/skills-sh-catalog.json.gz graph/communities.json graph/entity-overlays.jsonl
-python scripts/graph_artifact_guard.py prune
-```
-
-If a local Git integration gets interrupted while artifacts are dirty,
-`python scripts/graph_artifact_guard.py prune` removes prunable local LFS cache
-entries. It does not delete tracked graph files, rewrite history, or change the
-remote LFS store. Repo-wide `git prune --expire=now` is intentionally opt-in via
-`--include-git-prune` because it can discard unrelated dangling recovery objects.
-
-For a bulk skill refresh, update the existing shipped tarball through the
-release refresh path:
-
-```bash
-python src/import_skills_sh_catalog.py \
-  --from-catalog <skill-catalog.json.gz> \
-  --catalog-out <skill-catalog.json.gz> \
-  --wiki-tar graph/wiki-graph.tar.gz \
-  --update-wiki-tar
-```
-
-For a full local wiki repack, write the tarball to the sibling staged path,
-then promote that staged candidate after validation:
-
-```bash
-cd ~/.claude/skill-wiki
-tar --force-local -czf /path/to/ctx/graph/wiki-graph.tar.gz.staged \
-    --exclude='.trash' \
-    --exclude='__pycache__' \
-    --exclude='./raw' \
-    --exclude='./.embedding-cache' \
-    --exclude='./.ingest-checkpoint' \
-    --exclude='./.enrich-checkpoint' \
-    --exclude='./.ctx' \
-    --exclude='./graphify-out/graph.pickle' \
-    --exclude='*.original' \
-    --exclude='*.lock' \
-    .
-cd /path/to/ctx
-python -c "from pathlib import Path; from ctx.core.wiki.artifact_promotion import promote_staged_artifact; from import_skills_sh_catalog import _validate_wiki_tarball_candidate; promote_staged_artifact(Path('graph/wiki-graph.tar.gz.staged'), Path('graph/wiki-graph.tar.gz'), validate=_validate_wiki_tarball_candidate)"
-```
-
-The repack command above is for Git Bash/MSYS. In Linux/macOS shells omit
-`--force-local`; in PowerShell use `tar -czf` without `--force-local`.
-
-Both flows validate candidates before atomic promotion. Each promoted artifact
-gets a sibling `*.promotion.json` file with current, candidate, and `last_good`
-hashes for review or rollback. The graph, delta, communities, report, and
-export manifest are shipped together and carry the same export ID so validation
-can reject mixed or partially refreshed graph generations. Raw `.original`
-backups, transient `.lock` files, and `.ctx/` queue state must not appear in
-the shipped tarball.
-
-## Implementation Notes
-
-The graph is built by `ctx.core.wiki.wiki_graphify` and the `ctx-wiki-graphify`
-console script. Edges blend semantic similarity, explicit tag overlap,
-slug-token overlap, source overlap, direct links, quality, usage, type affinity,
-and graph-structure signals where available. The shipped default
-`graph.min_edge_weight` is `0.03`, chosen from artifact calibration because it
-keeps the current topology intact while recording the real shipped floor.
-
-`nashsu/llm_wiki` was reviewed for design ideas around persistent wiki
-contracts, queues, retrieval, and graph maintenance. ctx does not vendor that
-code in this MIT repository.
+# Knowledge Graph Artifacts
+
+This directory ships the pre-built ctx LLM-wiki and knowledge graph.
+
+Current snapshot:
+
+- **102,928 graph nodes**
+- **2,913,960 graph edges**
+- **52 Louvain communities**
+- **91,464 skill entity pages**; **89,465** have hydrated catalog bodies
+- **467 agent pages**
+- **10,790 MCP server pages**
+- **207 harness pages**
+- **89,465 hydrated `SKILL.md` bodies**
+- **28,612 long skill bodies converted through the micro-skill gate**
+
+The runtime recommendation paths use this graph in two ways:
+
+- Development recommendations return skills, agents, and MCP servers only.
+- Custom/API/local model onboarding recommends harnesses using the higher harness fit floor in `src/config.json`.
+
+## Files
+
+| File | Contents |
+|---|---|
+| `wiki-graph-runtime.tar.gz` | Fast install artifact used by default `ctx-init --graph`: `graphify-out/*`, the skill index, 207 harness pages, wiki index files, and Obsidian metadata needed for recommendations and harness dry-runs without expanding every entity page |
+| `wiki-graph.tar.gz` | Full LLM-wiki: entity pages, converted skill bodies, mirrored agent bodies, concept pages, `graphify-out/graph.json`, `graph-delta.json`, export manifest, communities, skill indexes, and Obsidian metadata |
+| Skill catalog gzip | Compressed skill index for the 89,465 body-backed skill entries shipped in the wiki |
+| `communities.json` | Current Louvain community export |
+| `entity-overlays.jsonl` | Release overlay for first-class entities added after the base graph export; installed beside `graphify-out/graph.json` by `ctx-init --graph` |
+| `graphify-out/dashboard-neighborhoods.sqlite3` inside both tarballs | Compact top-neighbor index used by `ctx-monitor` so `/api/graph/<slug>.json` does not cold-parse the 818 MiB graph JSON |
+| `viz-overview.html` | Plotly overview of the graph |
+| `viz-python.html` | Python-focused graph view |
+| `viz-security.html` | Security-focused graph view |
+| `viz-ai-agents.html` | AI-agent-focused graph view |
+| `sample-top60.html` | Interactive top-degree sample |
+
+Preview HTML files are generated from the shipped `graphify-out/graph.json`
+and embed the graph export ID in `<meta name="ctx-graph-export-id">`. Static
+PNG snapshots are intentionally not shipped because they can drift from the
+current tarball without an executable freshness check.
+
+## Runtime vs Full Wiki
+
+`ctx-init --graph` installs `wiki-graph-runtime.tar.gz` by default. That is the
+right path for recommendations and first-time installs because it avoids
+expanding hundreds of thousands of markdown files while still shipping the
+harness pages needed by `ctx-harness-install --dry-run`. Use
+`ctx-init --graph --graph-install-mode full` or manual full extraction when you
+want local wiki browsing, Obsidian, or the converted skill body tree.
+
+## What Is Inside `wiki-graph.tar.gz`
+
+- `entities/skills/` - all skill entity pages
+- `entities/agents/` - agent entity pages
+- `entities/mcp-servers/<shard>/` - sharded MCP server entity pages
+- `entities/harnesses/` - harness entity pages
+- `converted/` - installable skill bodies
+- `converted-agents/` - mirrored agent bodies
+- `concepts/` - community concept pages
+- `external-catalogs/` - machine-readable skill index, summary, and coverage metadata
+- `graphify-out/graph.json` - NetworkX node-link graph
+- `graphify-out/graph-delta.json` - delta export for the latest graph generation
+- `graphify-out/graph-export-manifest.json` - export manifest tying graph, delta, communities, and report to one generation
+- `graphify-out/communities.json` - community export
+- `SCHEMA.md`, `index.md`, `log.md`, `catalog.md` - wiki contract and indexes
+- `.obsidian/` - vault metadata for local graph browsing
+
+`SKILL.md.original` backups, transient `.lock` files, and `.ctx/` queue state
+are not shipped. Local micro-skill conversion may keep `.original` files for
+traceability, but the packaged tarball excludes them so users do not ingest raw
+long bodies after conversion.
+
+## Extract
+
+Default runtime install:
+
+```bash
+ctx-init --graph
+```
+
+Full wiki extraction:
+
+```bash
+mkdir -p ~/.claude/skill-wiki
+tar xzf graph/wiki-graph.tar.gz -C ~/.claude/skill-wiki/
+```
+
+On Windows PowerShell, use the built-in `tar.exe` without `--force-local`:
+
+```powershell
+New-Item -ItemType Directory -Force "$env:USERPROFILE\.claude\skill-wiki"
+tar -xzf graph\wiki-graph.tar.gz -C "$env:USERPROFILE\.claude\skill-wiki"
+```
+
+With Git Bash or MSYS tar, use `--force-local` only when the `-C` target is a
+drive-letter path:
+
+```bash
+tar --force-local xzf graph/wiki-graph.tar.gz -C C:/Users/<you>/.claude/skill-wiki/
+```
+
+## Validate
+
+```bash
+python src/validate_graph_artifacts.py --deep
+python src/update_repo_stats.py --check
+```
+
+For release-count validation, pin the exact snapshot numbers:
+
+```bash
+python src/validate_graph_artifacts.py --deep \
+  --expected-nodes 102928 \
+  --expected-edges 2913960 \
+  --expected-semantic-edges 1683193 \
+  --expected-harness-nodes 207 \
+  --expected-skill-pages 91464 \
+  --expected-agent-pages 467 \
+  --expected-mcp-pages 10790 \
+  --expected-harness-pages 207
+```
+
+Manual sanity checks:
+
+```bash
+tar -tzf graph/wiki-graph.tar.gz | grep 'graphify-out/graph.json'
+tar -tzf graph/wiki-graph.tar.gz | grep 'external-catalogs/.*/catalog.json'
+tar -tzf graph/wiki-graph.tar.gz | grep 'SKILL.md.original' && exit 1 || true
+tar -tzf graph/wiki-graph.tar.gz | grep '\.lock$' && exit 1 || true
+tar -tzf graph/wiki-graph.tar.gz | grep '^\./\.ctx/' && exit 1 || true
+```
+
+Windows PowerShell equivalent for the exclusion checks:
+
+```powershell
+tar -tzf graph/wiki-graph.tar.gz | Select-String 'SKILL.md.original'
+tar -tzf graph/wiki-graph.tar.gz | Select-String '\.lock$'
+tar -tzf graph/wiki-graph.tar.gz | Select-String '^\./\.ctx/'
+```
+
+The PowerShell commands should print nothing.
+
+## Rebuild
+
+After adding or updating skills, agents, MCP servers, or harnesses:
+
+```bash
+ctx-wiki-worker --wiki ~/.claude/skill-wiki --limit 1
+ctx-scan-repo --repo . --recommend
+```
+
+The worker path is the fast local update path. It validates the queued entity
+page, updates the wiki index, and attempts incremental ANN attach into
+`graphify-out/entity-overlays.jsonl` when the semantic vector index exists. It
+also queues the normal incremental graph export job, so a full rebuild remains
+the reconciliation path for release artifacts.
+
+If the worker reports that incremental attach was skipped because no vector
+index exists, build the exact portable index:
+
+```bash
+ctx-wiki-graphify \
+  --wiki-dir ~/.claude/skill-wiki \
+  --incremental \
+  --graph-only \
+  --semantic-vector-index numpy-flat
+```
+
+Then drain pending queue work again:
+
+```bash
+ctx-wiki-worker --wiki ~/.claude/skill-wiki
+```
+
+Before promoting an ANN backend or changed thresholds, run the shadow gate:
+
+```bash
+ctx-incremental-shadow \
+  --index-dir ~/.claude/skill-wiki/.embedding-cache/graph/vector-index \
+  --graph ~/.claude/skill-wiki/graphify-out/graph.json \
+  --sample-size 100 \
+  --min-overlap 0.85
+```
+
+It reports precision/recall, top-k agreement, score deltas, and bad examples;
+the release gate fails when recall at the largest requested top-k is below the
+overlap floor.
+
+For release artifact rebuilds:
+
+```bash
+python scripts/graph_artifact_guard.py park
+ctx-wiki-graphify
+python src/validate_graph_artifacts.py --deep
+python src/update_repo_stats.py --check
+```
+
+`park` sets Git's local `skip-worktree` bit for the heavyweight generated
+archives: `graph/wiki-graph.tar.gz`, `graph/wiki-graph-runtime.tar.gz`, and
+the compressed skill index. Keep them parked while graph/wiki generation,
+validation, dashboard smoke, and stats checks are still in progress. This
+prevents background Git integrations from repeatedly staging hundreds of
+megabytes through the Git LFS clean filter. When the release candidate is final,
+unpark and stage the artifacts exactly once:
+
+```bash
+python scripts/graph_artifact_guard.py unpark
+git add graph/wiki-graph.tar.gz graph/wiki-graph-runtime.tar.gz \
+  graph/skills-sh-catalog.json.gz graph/communities.json graph/entity-overlays.jsonl
+python scripts/graph_artifact_guard.py prune
+```
+
+If a local Git integration gets interrupted while artifacts are dirty,
+`python scripts/graph_artifact_guard.py prune` removes prunable local LFS cache
+entries. It does not delete tracked graph files, rewrite history, or change the
+remote LFS store. Repo-wide `git prune --expire=now` is intentionally opt-in via
+`--include-git-prune` because it can discard unrelated dangling recovery objects.
+
+For a bulk skill refresh, update the existing shipped tarball through the
+release refresh path:
+
+```bash
+python src/import_skills_sh_catalog.py \
+  --from-catalog <skill-catalog.json.gz> \
+  --catalog-out <skill-catalog.json.gz> \
+  --wiki-tar graph/wiki-graph.tar.gz \
+  --update-wiki-tar
+```
+
+For a full local wiki repack, write the tarball to the sibling staged path,
+then promote that staged candidate after validation:
+
+```bash
+cd ~/.claude/skill-wiki
+tar --force-local -czf /path/to/ctx/graph/wiki-graph.tar.gz.staged \
+    --exclude='.trash' \
+    --exclude='__pycache__' \
+    --exclude='./raw' \
+    --exclude='./.embedding-cache' \
+    --exclude='./.ingest-checkpoint' \
+    --exclude='./.enrich-checkpoint' \
+    --exclude='./.ctx' \
+    --exclude='./graphify-out/graph.pickle' \
+    --exclude='*.original' \
+    --exclude='*.lock' \
+    .
+cd /path/to/ctx
+python -c "from pathlib import Path; from ctx.core.wiki.artifact_promotion import promote_staged_artifact; from import_skills_sh_catalog import _validate_wiki_tarball_candidate; promote_staged_artifact(Path('graph/wiki-graph.tar.gz.staged'), Path('graph/wiki-graph.tar.gz'), validate=_validate_wiki_tarball_candidate)"
+```
+
+The repack command above is for Git Bash/MSYS. In Linux/macOS shells omit
+`--force-local`; in PowerShell use `tar -czf` without `--force-local`.
+
+Both flows validate candidates before atomic promotion. Each promoted artifact
+gets a sibling `*.promotion.json` file with current, candidate, and `last_good`
+hashes for review or rollback. The graph, delta, communities, report, and
+export manifest are shipped together and carry the same export ID so validation
+can reject mixed or partially refreshed graph generations. Raw `.original`
+backups, transient `.lock` files, and `.ctx/` queue state must not appear in
+the shipped tarball.
+
+## Implementation Notes
+
+The graph is built by `ctx.core.wiki.wiki_graphify` and the `ctx-wiki-graphify`
+console script. Edges blend semantic similarity, explicit tag overlap,
+slug-token overlap, source overlap, direct links, quality, usage, type affinity,
+and graph-structure signals where available. The shipped default
+`graph.min_edge_weight` is `0.03`, chosen from artifact calibration because it
+keeps the current topology intact while recording the real shipped floor.
+
+`nashsu/llm_wiki` was reviewed for design ideas around persistent wiki
+contracts, queues, retrieval, and graph maintenance. ctx does not vendor that
+code in this MIT repository.

imported-skills/julius-caveman/LICENSE

@@ -1,21 +1,21 @@
-MIT License
-
-Copyright (c) 2026 Julius Brussee
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+MIT License
+
+Copyright (c) 2026 Julius Brussee
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

imported-skills/julius-caveman/MANIFEST.json

@@ -1,123 +1,123 @@
-{
-  "upstream": "https://github.com/JuliusBrussee/caveman",
-  "upstream_revision": "63a91ecadbf4",
-  "license": "MIT",
-  "namespace": "julius-caveman",
-  "total": 10,
-  "skills": 7,
-  "agents": 3,
-  "entries": [
-    {
-      "name": "cavecrew",
-      "description": "Decision guide for delegating to caveman-style subagents. Tells the main thread WHEN to spawn `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit), or `cavecrew-reviewer` (diff review) instead of doing the work inline or using vanilla `Explore`. Subagent output is caveman-compressed so the tool-result injected back into main context is ~60% smaller \u2014 main context lasts longer across long sessions. Trigger: \"delegate to subagent\", \"use cavecrew\", \"spawn investigator/builder/reviewer\", \"save context\", \"compressed agent output\".",
-      "slug": "cavecrew",
-      "type": "skill",
-      "source_path": "skills/cavecrew/SKILL.md",
-      "support_files": [
-        "README.md"
-      ],
-      "lines": 82
-    },
-    {
-      "name": "caveman",
-      "description": "Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says \"caveman mode\", \"talk like caveman\", \"use caveman\", \"less tokens\", \"be brief\", or invokes /caveman. Also auto-triggers when token efficiency is requested.",
-      "slug": "caveman",
-      "type": "skill",
-      "source_path": "skills/caveman/SKILL.md",
-      "support_files": [
-        "README.md"
-      ],
-      "lines": 74
-    },
-    {
-      "name": "caveman-commit",
-      "description": "Ultra-compressed commit message generator. Cuts noise from commit messages while preserving intent and reasoning. Conventional Commits format. Subject \u226450 chars, body only when \"why\" isn't obvious. Use when user says \"write a commit\", \"commit message\", \"generate commit\", \"/commit\", or invokes /caveman-commit. Auto-triggers when staging changes.",
-      "slug": "caveman-commit",
-      "type": "skill",
-      "source_path": "skills/caveman-commit/SKILL.md",
-      "support_files": [
-        "README.md"
-      ],
-      "lines": 65
-    },
-    {
-      "name": "caveman-compress",
-      "description": "Compress natural language memory files (CLAUDE.md, todos, preferences) into caveman format to save input tokens. Preserves all technical substance, code, URLs, and structure. Compressed version overwrites the original file. Human-readable backup saved as FILE.original.md. Trigger: /caveman-compress FILEPATH or \"compress memory file\"",
-      "slug": "caveman-compress",
-      "type": "skill",
-      "source_path": "skills/caveman-compress/SKILL.md",
-      "support_files": [
-        "README.md",
-        "scripts/__init__.py",
-        "scripts/__main__.py",
-        "scripts/benchmark.py",
-        "scripts/cli.py",
-        "scripts/compress.py",
-        "scripts/detect.py",
-        "scripts/validate.py",
-        "SECURITY.md"
-      ],
-      "lines": 111
-    },
-    {
-      "name": "caveman-help",
-      "description": "Quick-reference card for all caveman modes, skills, and commands. One-shot display, not a persistent mode. Trigger: /caveman-help, \"caveman help\", \"what caveman commands\", \"how do I use caveman\".",
-      "slug": "caveman-help",
-      "type": "skill",
-      "source_path": "skills/caveman-help/SKILL.md",
-      "support_files": [
-        "README.md"
-      ],
-      "lines": 59
-    },
-    {
-      "name": "caveman-review",
-      "description": "Ultra-compressed code review comments. Cuts noise from PR feedback while preserving the actionable signal. Each comment is one line: location, problem, fix. Use when user says \"review this PR\", \"code review\", \"review the diff\", \"/review\", or invokes /caveman-review. Auto-triggers when reviewing pull requests.",
-      "slug": "caveman-review",
-      "type": "skill",
-      "source_path": "skills/caveman-review/SKILL.md",
-      "support_files": [
-        "README.md"
-      ],
-      "lines": 55
-    },
-    {
-      "name": "caveman-stats",
-      "description": "Show real token usage and estimated savings for the current session. Reads directly from the Claude Code session log \u2014 no AI estimation. Triggers on /caveman-stats. Output is injected by the mode-tracker hook; the model itself does not compute the numbers.",
-      "slug": "caveman-stats",
-      "type": "skill",
-      "source_path": "skills/caveman-stats/SKILL.md",
-      "support_files": [
-        "README.md"
-      ],
-      "lines": 10
-    },
-    {
-      "name": "cavecrew-builder",
-      "description": "Surgical 1-2 file edit. Typo fixes, single-function rewrites, mechanical renames, comment removal, format-preserving tweaks. Hard refuses 3+ file scope. Returns caveman diff receipt. Use when scope is bounded and obvious; do NOT use for new features, new files (unless asked), or cross-file refactors.",
-      "slug": "cavecrew-builder",
-      "type": "agent",
-      "source_path": "agents/cavecrew-builder.md",
-      "support_files": [],
-      "lines": 47
-    },
-    {
-      "name": "cavecrew-investigator",
-      "description": "Read-only code locator. Returns file:line table for \"where is X defined\", \"what calls Y\", \"list all uses of Z\", \"map this directory\". Output is caveman-compressed so the main thread eats ~60% fewer tokens than vanilla Explore. Refuses to suggest fixes.",
-      "slug": "cavecrew-investigator",
-      "type": "agent",
-      "source_path": "agents/cavecrew-investigator.md",
-      "support_files": [],
-      "lines": 57
-    },
-    {
-      "name": "cavecrew-reviewer",
-      "description": "Diff/branch/file reviewer. One line per finding, severity-tagged, no praise, no scope creep. Output format `path:line: <emoji> <severity>: <problem>. <fix>.` Use for \"review this PR\", \"review my diff\", \"audit this file\". Skips formatting nits unless they change meaning.",
-      "slug": "cavecrew-reviewer",
-      "type": "agent",
-      "source_path": "agents/cavecrew-reviewer.md",
-      "support_files": [],
-      "lines": 48
-    }
-  ]
-}
+{
+  "upstream": "https://github.com/JuliusBrussee/caveman",
+  "upstream_revision": "63a91ecadbf4",
+  "license": "MIT",
+  "namespace": "julius-caveman",
+  "total": 10,
+  "skills": 7,
+  "agents": 3,
+  "entries": [
+    {
+      "name": "cavecrew",
+      "description": "Decision guide for delegating to caveman-style subagents. Tells the main thread WHEN to spawn `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit), or `cavecrew-reviewer` (diff review) instead of doing the work inline or using vanilla `Explore`. Subagent output is caveman-compressed so the tool-result injected back into main context is ~60% smaller \u2014 main context lasts longer across long sessions. Trigger: \"delegate to subagent\", \"use cavecrew\", \"spawn investigator/builder/reviewer\", \"save context\", \"compressed agent output\".",
+      "slug": "cavecrew",
+      "type": "skill",
+      "source_path": "skills/cavecrew/SKILL.md",
+      "support_files": [
+        "README.md"
+      ],
+      "lines": 82
+    },
+    {
+      "name": "caveman",
+      "description": "Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says \"caveman mode\", \"talk like caveman\", \"use caveman\", \"less tokens\", \"be brief\", or invokes /caveman. Also auto-triggers when token efficiency is requested.",
+      "slug": "caveman",
+      "type": "skill",
+      "source_path": "skills/caveman/SKILL.md",
+      "support_files": [
+        "README.md"
+      ],
+      "lines": 74
+    },
+    {
+      "name": "caveman-commit",
+      "description": "Ultra-compressed commit message generator. Cuts noise from commit messages while preserving intent and reasoning. Conventional Commits format. Subject \u226450 chars, body only when \"why\" isn't obvious. Use when user says \"write a commit\", \"commit message\", \"generate commit\", \"/commit\", or invokes /caveman-commit. Auto-triggers when staging changes.",
+      "slug": "caveman-commit",
+      "type": "skill",
+      "source_path": "skills/caveman-commit/SKILL.md",
+      "support_files": [
+        "README.md"
+      ],
+      "lines": 65
+    },
+    {
+      "name": "caveman-compress",
+      "description": "Compress natural language memory files (CLAUDE.md, todos, preferences) into caveman format to save input tokens. Preserves all technical substance, code, URLs, and structure. Compressed version overwrites the original file. Human-readable backup saved as FILE.original.md. Trigger: /caveman-compress FILEPATH or \"compress memory file\"",
+      "slug": "caveman-compress",
+      "type": "skill",
+      "source_path": "skills/caveman-compress/SKILL.md",
+      "support_files": [
+        "README.md",
+        "scripts/__init__.py",
+        "scripts/__main__.py",
+        "scripts/benchmark.py",
+        "scripts/cli.py",
+        "scripts/compress.py",
+        "scripts/detect.py",
+        "scripts/validate.py",
+        "SECURITY.md"
+      ],
+      "lines": 111
+    },
+    {
+      "name": "caveman-help",
+      "description": "Quick-reference card for all caveman modes, skills, and commands. One-shot display, not a persistent mode. Trigger: /caveman-help, \"caveman help\", \"what caveman commands\", \"how do I use caveman\".",
+      "slug": "caveman-help",
+      "type": "skill",
+      "source_path": "skills/caveman-help/SKILL.md",
+      "support_files": [
+        "README.md"
+      ],
+      "lines": 59
+    },
+    {
+      "name": "caveman-review",
+      "description": "Ultra-compressed code review comments. Cuts noise from PR feedback while preserving the actionable signal. Each comment is one line: location, problem, fix. Use when user says \"review this PR\", \"code review\", \"review the diff\", \"/review\", or invokes /caveman-review. Auto-triggers when reviewing pull requests.",
+      "slug": "caveman-review",
+      "type": "skill",
+      "source_path": "skills/caveman-review/SKILL.md",
+      "support_files": [
+        "README.md"
+      ],
+      "lines": 55
+    },
+    {
+      "name": "caveman-stats",
+      "description": "Show real token usage and estimated savings for the current session. Reads directly from the Claude Code session log \u2014 no AI estimation. Triggers on /caveman-stats. Output is injected by the mode-tracker hook; the model itself does not compute the numbers.",
+      "slug": "caveman-stats",
+      "type": "skill",
+      "source_path": "skills/caveman-stats/SKILL.md",
+      "support_files": [
+        "README.md"
+      ],
+      "lines": 10
+    },
+    {
+      "name": "cavecrew-builder",
+      "description": "Surgical 1-2 file edit. Typo fixes, single-function rewrites, mechanical renames, comment removal, format-preserving tweaks. Hard refuses 3+ file scope. Returns caveman diff receipt. Use when scope is bounded and obvious; do NOT use for new features, new files (unless asked), or cross-file refactors.",
+      "slug": "cavecrew-builder",
+      "type": "agent",
+      "source_path": "agents/cavecrew-builder.md",
+      "support_files": [],
+      "lines": 47
+    },
+    {
+      "name": "cavecrew-investigator",
+      "description": "Read-only code locator. Returns file:line table for \"where is X defined\", \"what calls Y\", \"list all uses of Z\", \"map this directory\". Output is caveman-compressed so the main thread eats ~60% fewer tokens than vanilla Explore. Refuses to suggest fixes.",
+      "slug": "cavecrew-investigator",
+      "type": "agent",
+      "source_path": "agents/cavecrew-investigator.md",
+      "support_files": [],
+      "lines": 57
+    },
+    {
+      "name": "cavecrew-reviewer",
+      "description": "Diff/branch/file reviewer. One line per finding, severity-tagged, no praise, no scope creep. Output format `path:line: <emoji> <severity>: <problem>. <fix>.` Use for \"review this PR\", \"review my diff\", \"audit this file\". Skips formatting nits unless they change meaning.",
+      "slug": "cavecrew-reviewer",
+      "type": "agent",
+      "source_path": "agents/cavecrew-reviewer.md",
+      "support_files": [],
+      "lines": 48
+    }
+  ]
+}

imported-skills/julius-caveman/agents/cavecrew-builder.md

@@ -1,47 +1,47 @@
----
-name: cavecrew-builder
-description: >
-  Surgical 1-2 file edit. Typo fixes, single-function rewrites, mechanical
-  renames, comment removal, format-preserving tweaks. Hard refuses 3+ file
-  scope. Returns caveman diff receipt. Use when scope is bounded and
-  obvious; do NOT use for new features, new files (unless asked), or
-  cross-file refactors.
-tools: [Read, Edit, Write, Grep, Glob]
----
-
-Caveman-ultra. Drop articles/filler. Code/paths exact, backticked. No narration.
-
-## Scope
-
-1 file ideal. 2 OK. 3+ → refuse.
-Edit existing only (new file iff user asked).
-No new abstractions. No drive-by refactors. No comment additions.
-No `Bash` available — cannot shell out, cannot push, cannot delete.
-
-## Workflow
-
-1. `Read` target(s). Never edit blind.
-2. `Edit` smallest diff that work.
-3. Re-`Read` to verify.
-4. Return receipt.
-
-## Output (receipt)
-
-```
-<path:line-range> — <change ≤10 words>.
-<path:line-range> — <change ≤10 words>.
-verified: <re-read OK | mismatch @ path:line>.
-```
-
-Diff is the artifact. Receipt is the proof. No exploration story.
-
-## Refusals (terminal lines)
-
-3+ files → `too-big. split: <n one-line tasks>.`
-Destructive needed → `needs-confirm. op: <command>.`
-Spec ambiguous → `ambiguous. ask: <one question>.`
-Tests fail post-edit, can't fix in scope → `regressed. revert path:line. cause: <fragment>.`
-
-## Auto-clarity
-
-Security or destructive paths → write normal English warning, then resume caveman.
+---
+name: cavecrew-builder
+description: >
+  Surgical 1-2 file edit. Typo fixes, single-function rewrites, mechanical
+  renames, comment removal, format-preserving tweaks. Hard refuses 3+ file
+  scope. Returns caveman diff receipt. Use when scope is bounded and
+  obvious; do NOT use for new features, new files (unless asked), or
+  cross-file refactors.
+tools: [Read, Edit, Write, Grep, Glob]
+---
+
+Caveman-ultra. Drop articles/filler. Code/paths exact, backticked. No narration.
+
+## Scope
+
+1 file ideal. 2 OK. 3+ → refuse.
+Edit existing only (new file iff user asked).
+No new abstractions. No drive-by refactors. No comment additions.
+No `Bash` available — cannot shell out, cannot push, cannot delete.
+
+## Workflow
+
+1. `Read` target(s). Never edit blind.
+2. `Edit` smallest diff that work.
+3. Re-`Read` to verify.
+4. Return receipt.
+
+## Output (receipt)
+
+```
+<path:line-range> — <change ≤10 words>.
+<path:line-range> — <change ≤10 words>.
+verified: <re-read OK | mismatch @ path:line>.
+```
+
+Diff is the artifact. Receipt is the proof. No exploration story.
+
+## Refusals (terminal lines)
+
+3+ files → `too-big. split: <n one-line tasks>.`
+Destructive needed → `needs-confirm. op: <command>.`
+Spec ambiguous → `ambiguous. ask: <one question>.`
+Tests fail post-edit, can't fix in scope → `regressed. revert path:line. cause: <fragment>.`
+
+## Auto-clarity
+
+Security or destructive paths → write normal English warning, then resume caveman.

imported-skills/julius-caveman/agents/cavecrew-investigator.md

@@ -1,57 +1,57 @@
----
-name: cavecrew-investigator
-description: >
-  Read-only code locator. Returns file:line table for "where is X defined",
-  "what calls Y", "list all uses of Z", "map this directory". Output is
-  caveman-compressed so the main thread eats ~60% fewer tokens than
-  vanilla Explore. Refuses to suggest fixes.
-tools: [Read, Grep, Glob, Bash]
-model: haiku
----
-
-Caveman-ultra. Drop articles/filler/hedging. Code/symbols/paths exact, backticked. Lead with answer.
-
-## Job
-
-Locate. Report. Stop. Never edit, never propose fix.
-
-## Output
-
-```
-<path:line> — `<symbol>` — <≤6 word note>
-<path:line> — `<symbol>` — <≤6 word note>
-```
-
-Group with one-word header when 3+ rows: `Defs:` / `Refs:` / `Callers:` / `Tests:` / `Imports:` / `Sites:`.
-Single hit → one line, no header.
-Zero hits → `No match.`
-Last line → totals: `2 defs, 5 refs.` (omit if 0 or 1).
-
-## Tools
-
-`Grep` for symbols/strings. `Glob` for paths. `Read` only specific ranges. `Bash` for `git log -S`/`git grep`/`find` when faster.
-
-## Refusals
-
-Asked to fix → `Read-only. Spawn cavecrew-builder.`
-Asked to design → `Read-only. Spawn cavecrew-builder or use main thread.`
-
-## Auto-clarity
-
-Security warnings, destructive ops → write normal English. Resume after.
-
-## Example
-
-Q: "where symlink-safe flag write?"
-
-```
-Defs:
-- hooks/caveman-config.js:81 — `safeWriteFlag` — atomic write w/ O_NOFOLLOW
-- hooks/caveman-config.js:160 — `readFlag` — paired reader
-Callers:
-- hooks/caveman-mode-tracker.js:33,87
-- hooks/caveman-activate.js:40
-Tests:
-- tests/test_symlink_flag.js — 12 cases
-2 defs, 3 callers, 1 test file.
-```
+---
+name: cavecrew-investigator
+description: >
+  Read-only code locator. Returns file:line table for "where is X defined",
+  "what calls Y", "list all uses of Z", "map this directory". Output is
+  caveman-compressed so the main thread eats ~60% fewer tokens than
+  vanilla Explore. Refuses to suggest fixes.
+tools: [Read, Grep, Glob, Bash]
+model: haiku
+---
+
+Caveman-ultra. Drop articles/filler/hedging. Code/symbols/paths exact, backticked. Lead with answer.
+
+## Job
+
+Locate. Report. Stop. Never edit, never propose fix.
+
+## Output
+
+```
+<path:line> — `<symbol>` — <≤6 word note>
+<path:line> — `<symbol>` — <≤6 word note>
+```
+
+Group with one-word header when 3+ rows: `Defs:` / `Refs:` / `Callers:` / `Tests:` / `Imports:` / `Sites:`.
+Single hit → one line, no header.
+Zero hits → `No match.`
+Last line → totals: `2 defs, 5 refs.` (omit if 0 or 1).
+
+## Tools
+
+`Grep` for symbols/strings. `Glob` for paths. `Read` only specific ranges. `Bash` for `git log -S`/`git grep`/`find` when faster.
+
+## Refusals
+
+Asked to fix → `Read-only. Spawn cavecrew-builder.`
+Asked to design → `Read-only. Spawn cavecrew-builder or use main thread.`
+
+## Auto-clarity
+
+Security warnings, destructive ops → write normal English. Resume after.
+
+## Example
+
+Q: "where symlink-safe flag write?"
+
+```
+Defs:
+- hooks/caveman-config.js:81 — `safeWriteFlag` — atomic write w/ O_NOFOLLOW
+- hooks/caveman-config.js:160 — `readFlag` — paired reader
+Callers:
+- hooks/caveman-mode-tracker.js:33,87
+- hooks/caveman-activate.js:40
+Tests:
+- tests/test_symlink_flag.js — 12 cases
+2 defs, 3 callers, 1 test file.
+```

imported-skills/julius-caveman/agents/cavecrew-reviewer.md

@@ -1,48 +1,48 @@
----
-name: cavecrew-reviewer
-description: >
-  Diff/branch/file reviewer. One line per finding, severity-tagged, no praise,
-  no scope creep. Output format `path:line: <emoji> <severity>: <problem>. <fix>.`
-  Use for "review this PR", "review my diff", "audit this file". Skips
-  formatting nits unless they change meaning.
-tools: [Read, Grep, Bash]
-model: haiku
----
-
-Caveman-ultra. Findings only. No "looks good", no "I'd suggest", no preamble.
-
-## Severity
-
-| Emoji | Tier | Use for |
-|---|---|---|
-| 🔴 | bug | Wrong output, crash, security hole, data loss |
-| 🟡 | risk | Edge case, race, leak, perf cliff, missing guard |
-| 🔵 | nit | Style, naming, micro-perf — emit only if user asked thorough |
-| ❓ | question | Need author intent before judging |
-
-## Output
-
-```
-path/to/file.ts:42: 🔴 bug: token expiry uses `<` not `<=`. Off-by-one allows expired tokens 1 tick.
-path/to/file.ts:118: 🟡 risk: pool not closed on error path. Add `try/finally`.
-src/utils.ts:7: ❓ question: why duplicate `.trim()` here?
-totals: 1🔴 1🟡 1❓
-```
-
-Zero findings → `No issues.`
-File order, ascending line numbers within file.
-
-## Boundaries
-
-- Review only what's in front of you. No "while we're here".
-- No big-refactor proposals.
-- Need more context → append `(see L<n> in <file>)`. Don't guess.
-- Formatting nits skipped unless they change meaning.
-
-## Tools
-
-`Bash` only for `git diff`/`git log -p`/`git show`. No mutating commands.
-
-## Auto-clarity
-
-Security findings → state risk in plain English first sentence, then caveman fix line.
+---
+name: cavecrew-reviewer
+description: >
+  Diff/branch/file reviewer. One line per finding, severity-tagged, no praise,
+  no scope creep. Output format `path:line: <emoji> <severity>: <problem>. <fix>.`
+  Use for "review this PR", "review my diff", "audit this file". Skips
+  formatting nits unless they change meaning.
+tools: [Read, Grep, Bash]
+model: haiku
+---
+
+Caveman-ultra. Findings only. No "looks good", no "I'd suggest", no preamble.
+
+## Severity
+
+| Emoji | Tier | Use for |
+|---|---|---|
+| 🔴 | bug | Wrong output, crash, security hole, data loss |
+| 🟡 | risk | Edge case, race, leak, perf cliff, missing guard |
+| 🔵 | nit | Style, naming, micro-perf — emit only if user asked thorough |
+| ❓ | question | Need author intent before judging |
+
+## Output
+
+```
+path/to/file.ts:42: 🔴 bug: token expiry uses `<` not `<=`. Off-by-one allows expired tokens 1 tick.
+path/to/file.ts:118: 🟡 risk: pool not closed on error path. Add `try/finally`.
+src/utils.ts:7: ❓ question: why duplicate `.trim()` here?
+totals: 1🔴 1🟡 1❓
+```
+
+Zero findings → `No issues.`
+File order, ascending line numbers within file.
+
+## Boundaries
+
+- Review only what's in front of you. No "while we're here".
+- No big-refactor proposals.
+- Need more context → append `(see L<n> in <file>)`. Don't guess.
+- Formatting nits skipped unless they change meaning.
+
+## Tools
+
+`Bash` only for `git diff`/`git log -p`/`git show`. No mutating commands.
+
+## Auto-clarity
+
+Security findings → state risk in plain English first sentence, then caveman fix line.

imported-skills/julius-caveman/skills/cavecrew/README.md

@@ -1,41 +1,41 @@
-# cavecrew
-
-Decision guide. When to delegate to caveman subagents instead of doing the work inline.
-
-## What it does
-
-Tells the main thread when to spawn a caveman-style subagent versus the vanilla equivalent. The win: subagent tool-results inject back into main context verbatim, and caveman output is roughly 1/3 the size of vanilla prose. Across 20 delegations in one session, that is the difference between context exhaustion and finishing the task.
-
-Three subagents:
-
-| Subagent | Job | Use when |
-|----------|-----|----------|
-| `cavecrew-investigator` | Locate code (read-only) | "Where is X defined / what calls Y / list uses of Z" |
-| `cavecrew-builder` | Surgical edit, 1-2 files | Scope is obvious, ≤2 files. Refuses 3+ file scope. |
-| `cavecrew-reviewer` | Diff/file review | One-line findings with severity emoji |
-
-Use vanilla `Explore` or `Code Reviewer` when you want prose, architecture commentary, or rationale. Use main thread directly for one-line answers and 3+ file refactors.
-
-This skill is a decision guide, not a slash command. It activates when the conversation mentions delegation.
-
-## How to invoke
-
-Triggers on phrases like "delegate to subagent", "use cavecrew", "spawn investigator", "save context", "compressed agent output".
-
-## Example chaining
-
-Locate → fix → verify (most common):
-
-1. `cavecrew-investigator` returns site list (`path:line — symbol — note`)
-2. Main thread picks 1-2 sites, hands paths to `cavecrew-builder`
-3. `cavecrew-reviewer` audits the resulting diff
-
-Parallel scout: spawn 2-3 `cavecrew-investigator` calls in one message with different angles (defs, callers, tests). Aggregate in main.
-
-## See also
-
-- [`SKILL.md`](./SKILL.md) — full decision matrix and output contracts
-- [`agents/cavecrew-investigator.md`](../../agents/cavecrew-investigator.md)
-- [`agents/cavecrew-builder.md`](../../agents/cavecrew-builder.md)
-- [`agents/cavecrew-reviewer.md`](../../agents/cavecrew-reviewer.md)
-- [Caveman README](../../README.md) — repo overview
+# cavecrew
+
+Decision guide. When to delegate to caveman subagents instead of doing the work inline.
+
+## What it does
+
+Tells the main thread when to spawn a caveman-style subagent versus the vanilla equivalent. The win: subagent tool-results inject back into main context verbatim, and caveman output is roughly 1/3 the size of vanilla prose. Across 20 delegations in one session, that is the difference between context exhaustion and finishing the task.
+
+Three subagents:
+
+| Subagent | Job | Use when |
+|----------|-----|----------|
+| `cavecrew-investigator` | Locate code (read-only) | "Where is X defined / what calls Y / list uses of Z" |
+| `cavecrew-builder` | Surgical edit, 1-2 files | Scope is obvious, ≤2 files. Refuses 3+ file scope. |
+| `cavecrew-reviewer` | Diff/file review | One-line findings with severity emoji |
+
+Use vanilla `Explore` or `Code Reviewer` when you want prose, architecture commentary, or rationale. Use main thread directly for one-line answers and 3+ file refactors.
+
+This skill is a decision guide, not a slash command. It activates when the conversation mentions delegation.
+
+## How to invoke
+
+Triggers on phrases like "delegate to subagent", "use cavecrew", "spawn investigator", "save context", "compressed agent output".
+
+## Example chaining
+
+Locate → fix → verify (most common):
+
+1. `cavecrew-investigator` returns site list (`path:line — symbol — note`)
+2. Main thread picks 1-2 sites, hands paths to `cavecrew-builder`
+3. `cavecrew-reviewer` audits the resulting diff
+
+Parallel scout: spawn 2-3 `cavecrew-investigator` calls in one message with different angles (defs, callers, tests). Aggregate in main.
+
+## See also
+
+- [`SKILL.md`](./SKILL.md) — full decision matrix and output contracts
+- [`agents/cavecrew-investigator.md`](../../agents/cavecrew-investigator.md)
+- [`agents/cavecrew-builder.md`](../../agents/cavecrew-builder.md)
+- [`agents/cavecrew-reviewer.md`](../../agents/cavecrew-reviewer.md)
+- [Caveman README](../../README.md) — repo overview

imported-skills/julius-caveman/skills/cavecrew/SKILL.md

@@ -1,82 +1,82 @@
----
-name: cavecrew
-description: >
-  Decision guide for delegating to caveman-style subagents. Tells the main
-  thread WHEN to spawn `cavecrew-investigator` (locate code), `cavecrew-builder`
-  (1-2 file edit), or `cavecrew-reviewer` (diff review) instead of doing the
-  work inline or using vanilla `Explore`. Subagent output is caveman-compressed
-  so the tool-result injected back into main context is ~60% smaller — main
-  context lasts longer across long sessions.
-  Trigger: "delegate to subagent", "use cavecrew", "spawn investigator/builder/reviewer",
-  "save context", "compressed agent output".
----
-
-Cavecrew = three subagent presets that emit caveman output. Same job as Anthropic defaults (`Explore`, edit-style agents, reviewer); difference is the tool-result they return is compressed, so main context shrinks per delegation.
-
-## When to use cavecrew vs alternatives
-
-| Task | Use |
-|---|---|
-| "Where is X defined / what calls Y / list uses of Z" | `cavecrew-investigator` |
-| Same but you also want suggestions/architecture commentary | `Explore` (vanilla) |
-| Surgical edit, ≤2 files, scope obvious | `cavecrew-builder` |
-| New feature / 3+ files / cross-cutting refactor | Main thread or `feature-dev:code-architect` |
-| Review diff, branch, or file for bugs | `cavecrew-reviewer` |
-| Deep code review with rationale + alternatives | `Code Reviewer` (vanilla) |
-| One-line answer you already know | Main thread, no subagent |
-
-Rule of thumb: **if you'd want the subagent's output in 1/3 the tokens, pick cavecrew. If you'd want prose, pick vanilla.**
-
-## Why this exists (the real win)
-
-Subagent tool results get injected into main context verbatim. A vanilla `Explore` that returns 2k tokens of prose costs 2k tokens of main-context budget every time. The same finding from `cavecrew-investigator` returns ~700 tokens. Across 20 delegations in one session that's the difference between context exhaustion and finishing the task.
-
-## Output contracts
-
-What main thread can rely on per agent:
-
-**`cavecrew-investigator`**
-```
-<Header>:
-- path:line — `symbol` — short note
-totals: <counts>.
-```
-Or `No match.` Always file-path-first, line-number-attached, backticked symbols. Safe to grep with `path:\d+`.
-
-**`cavecrew-builder`**
-```
-<path:line-range> — <change ≤10 words>.
-verified: <re-read OK | mismatch @ path:line>.
-```
-Or one of: `too-big.` / `needs-confirm.` / `ambiguous.` / `regressed.` (terminal first token).
-
-**`cavecrew-reviewer`**
-```
-path:line: <emoji> <severity>: <problem>. <fix>.
-totals: N🔴 N🟡 N🔵 N❓
-```
-Or `No issues.` Findings sorted file → line ascending.
-
-## Chaining patterns
-
-**Locate → fix → verify** (most common):
-1. `cavecrew-investigator` returns site list.
-2. Main thread picks 1-2 sites, hands paths to `cavecrew-builder`.
-3. `cavecrew-reviewer` audits the diff.
-
-**Parallel scout** (when investigation is broad):
-Spawn 2-3 `cavecrew-investigator` calls in one message (different angles: defs vs callers vs tests). Aggregate in main thread.
-
-**Single-shot edit** (when site is already known):
-Skip investigator. Hand exact path:line to `cavecrew-builder` directly.
-
-## What NOT to do
-
-- Don't use `cavecrew-builder` when you don't already know the file. Spawn investigator first or main thread will eat tokens passing context.
-- Don't chain `cavecrew-investigator → cavecrew-builder` for a 5-file refactor. Builder will return `too-big.` and you'll have wasted a turn.
-- Don't ask `cavecrew-reviewer` for "general feedback" — it returns findings only, no architecture opinions. Use `Code Reviewer` for that.
-- Don't expect prose. Cavecrew output is structured, sometimes terse to the point of cryptic. If a human will read it directly, paraphrase.
-
-## Auto-clarity (inherited)
-
-Subagents drop caveman → normal English for security warnings, irreversible-action confirmations, and any output where fragment ambiguity could be misread. Resume caveman after.
+---
+name: cavecrew
+description: >
+  Decision guide for delegating to caveman-style subagents. Tells the main
+  thread WHEN to spawn `cavecrew-investigator` (locate code), `cavecrew-builder`
+  (1-2 file edit), or `cavecrew-reviewer` (diff review) instead of doing the
+  work inline or using vanilla `Explore`. Subagent output is caveman-compressed
+  so the tool-result injected back into main context is ~60% smaller — main
+  context lasts longer across long sessions.
+  Trigger: "delegate to subagent", "use cavecrew", "spawn investigator/builder/reviewer",
+  "save context", "compressed agent output".
+---
+
+Cavecrew = three subagent presets that emit caveman output. Same job as Anthropic defaults (`Explore`, edit-style agents, reviewer); difference is the tool-result they return is compressed, so main context shrinks per delegation.
+
+## When to use cavecrew vs alternatives
+
+| Task | Use |
+|---|---|
+| "Where is X defined / what calls Y / list uses of Z" | `cavecrew-investigator` |
+| Same but you also want suggestions/architecture commentary | `Explore` (vanilla) |
+| Surgical edit, ≤2 files, scope obvious | `cavecrew-builder` |
+| New feature / 3+ files / cross-cutting refactor | Main thread or `feature-dev:code-architect` |
+| Review diff, branch, or file for bugs | `cavecrew-reviewer` |
+| Deep code review with rationale + alternatives | `Code Reviewer` (vanilla) |
+| One-line answer you already know | Main thread, no subagent |
+
+Rule of thumb: **if you'd want the subagent's output in 1/3 the tokens, pick cavecrew. If you'd want prose, pick vanilla.**
+
+## Why this exists (the real win)
+
+Subagent tool results get injected into main context verbatim. A vanilla `Explore` that returns 2k tokens of prose costs 2k tokens of main-context budget every time. The same finding from `cavecrew-investigator` returns ~700 tokens. Across 20 delegations in one session that's the difference between context exhaustion and finishing the task.
+
+## Output contracts
+
+What main thread can rely on per agent:
+
+**`cavecrew-investigator`**
+```
+<Header>:
+- path:line — `symbol` — short note
+totals: <counts>.
+```
+Or `No match.` Always file-path-first, line-number-attached, backticked symbols. Safe to grep with `path:\d+`.
+
+**`cavecrew-builder`**
+```
+<path:line-range> — <change ≤10 words>.
+verified: <re-read OK | mismatch @ path:line>.
+```
+Or one of: `too-big.` / `needs-confirm.` / `ambiguous.` / `regressed.` (terminal first token).
+
+**`cavecrew-reviewer`**
+```
+path:line: <emoji> <severity>: <problem>. <fix>.
+totals: N🔴 N🟡 N🔵 N❓
+```
+Or `No issues.` Findings sorted file → line ascending.
+
+## Chaining patterns
+
+**Locate → fix → verify** (most common):
+1. `cavecrew-investigator` returns site list.
+2. Main thread picks 1-2 sites, hands paths to `cavecrew-builder`.
+3. `cavecrew-reviewer` audits the diff.
+
+**Parallel scout** (when investigation is broad):
+Spawn 2-3 `cavecrew-investigator` calls in one message (different angles: defs vs callers vs tests). Aggregate in main thread.
+
+**Single-shot edit** (when site is already known):
+Skip investigator. Hand exact path:line to `cavecrew-builder` directly.
+
+## What NOT to do
+
+- Don't use `cavecrew-builder` when you don't already know the file. Spawn investigator first or main thread will eat tokens passing context.
+- Don't chain `cavecrew-investigator → cavecrew-builder` for a 5-file refactor. Builder will return `too-big.` and you'll have wasted a turn.
+- Don't ask `cavecrew-reviewer` for "general feedback" — it returns findings only, no architecture opinions. Use `Code Reviewer` for that.
+- Don't expect prose. Cavecrew output is structured, sometimes terse to the point of cryptic. If a human will read it directly, paraphrase.
+
+## Auto-clarity (inherited)
+
+Subagents drop caveman → normal English for security warnings, irreversible-action confirmations, and any output where fragment ambiguity could be misread. Resume caveman after.

imported-skills/julius-caveman/skills/caveman-commit/README.md

@@ -1,44 +1,44 @@
-# caveman-commit
-
-Terse Conventional Commits. Why over what.
-
-## What it does
-
-Generates commit messages in Conventional Commits format. Subject ≤50 chars, hard cap 72. Imperative mood. Body only when the *why* is non-obvious or there are breaking changes. No AI attribution, no "this commit does X", no emoji unless the project uses them. Body always required for breaking changes, security fixes, data migrations, and reverts — future debuggers need the context.
-
-Outputs only the message. Does not stage, commit, or amend.
-
-## How to invoke
-
-```
-/caveman-commit
-```
-
-Also triggers on phrases like "write a commit", "commit message", "generate commit".
-
-## Example output
-
-Diff: new endpoint for user profile.
-
-```
-feat(api): add GET /users/:id/profile
-
-Mobile client needs profile data without the full user payload
-to reduce LTE bandwidth on cold-launch screens.
-
-Closes #128
-```
-
-Diff: breaking API rename.
-
-```
-feat(api)!: rename /v1/orders to /v1/checkout
-
-BREAKING CHANGE: clients on /v1/orders must migrate to /v1/checkout
-before 2026-06-01. Old route returns 410 after that date.
-```
-
-## See also
-
-- [`SKILL.md`](./SKILL.md) — full LLM-facing instructions
-- [Caveman README](../../README.md) — repo overview
+# caveman-commit
+
+Terse Conventional Commits. Why over what.
+
+## What it does
+
+Generates commit messages in Conventional Commits format. Subject ≤50 chars, hard cap 72. Imperative mood. Body only when the *why* is non-obvious or there are breaking changes. No AI attribution, no "this commit does X", no emoji unless the project uses them. Body always required for breaking changes, security fixes, data migrations, and reverts — future debuggers need the context.
+
+Outputs only the message. Does not stage, commit, or amend.
+
+## How to invoke
+
+```
+/caveman-commit
+```
+
+Also triggers on phrases like "write a commit", "commit message", "generate commit".
+
+## Example output
+
+Diff: new endpoint for user profile.
+
+```
+feat(api): add GET /users/:id/profile
+
+Mobile client needs profile data without the full user payload
+to reduce LTE bandwidth on cold-launch screens.
+
+Closes #128
+```
+
+Diff: breaking API rename.
+
+```
+feat(api)!: rename /v1/orders to /v1/checkout
+
+BREAKING CHANGE: clients on /v1/orders must migrate to /v1/checkout
+before 2026-06-01. Old route returns 410 after that date.
+```
+
+## See also
+
+- [`SKILL.md`](./SKILL.md) — full LLM-facing instructions
+- [Caveman README](../../README.md) — repo overview

imported-skills/julius-caveman/skills/caveman-commit/SKILL.md

@@ -1,65 +1,65 @@
----
-name: caveman-commit
-description: >
-  Ultra-compressed commit message generator. Cuts noise from commit messages while preserving
-  intent and reasoning. Conventional Commits format. Subject ≤50 chars, body only when "why"
-  isn't obvious. Use when user says "write a commit", "commit message", "generate commit",
-  "/commit", or invokes /caveman-commit. Auto-triggers when staging changes.
----
-
-Write commit messages terse and exact. Conventional Commits format. No fluff. Why over what.
-
-## Rules
-
-**Subject line:**
-- `<type>(<scope>): <imperative summary>` — `<scope>` optional
-- Types: `feat`, `fix`, `refactor`, `perf`, `docs`, `test`, `chore`, `build`, `ci`, `style`, `revert`
-- Imperative mood: "add", "fix", "remove" — not "added", "adds", "adding"
-- ≤50 chars when possible, hard cap 72
-- No trailing period
-- Match project convention for capitalization after the colon
-
-**Body (only if needed):**
-- Skip entirely when subject is self-explanatory
-- Add body only for: non-obvious *why*, breaking changes, migration notes, linked issues
-- Wrap at 72 chars
-- Bullets `-` not `*`
-- Reference issues/PRs at end: `Closes #42`, `Refs #17`
-
-**What NEVER goes in:**
-- "This commit does X", "I", "we", "now", "currently" — the diff says what
-- "As requested by..." — use Co-authored-by trailer
-- "Generated with Claude Code" or any AI attribution
-- Emoji (unless project convention requires)
-- Restating the file name when scope already says it
-
-## Examples
-
-Diff: new endpoint for user profile with body explaining the why
-- ❌ "feat: add a new endpoint to get user profile information from the database"
-- ✅
-  ```
-  feat(api): add GET /users/:id/profile
-
-  Mobile client needs profile data without the full user payload
-  to reduce LTE bandwidth on cold-launch screens.
-
-  Closes #128
-  ```
-
-Diff: breaking API change
-- ✅
-  ```
-  feat(api)!: rename /v1/orders to /v1/checkout
-
-  BREAKING CHANGE: clients on /v1/orders must migrate to /v1/checkout
-  before 2026-06-01. Old route returns 410 after that date.
-  ```
-
-## Auto-Clarity
-
-Always include body for: breaking changes, security fixes, data migrations, anything reverting a prior commit. Never compress these into subject-only — future debuggers need the context.
-
-## Boundaries
-
+---
+name: caveman-commit
+description: >
+  Ultra-compressed commit message generator. Cuts noise from commit messages while preserving
+  intent and reasoning. Conventional Commits format. Subject ≤50 chars, body only when "why"
+  isn't obvious. Use when user says "write a commit", "commit message", "generate commit",
+  "/commit", or invokes /caveman-commit. Auto-triggers when staging changes.
+---
+
+Write commit messages terse and exact. Conventional Commits format. No fluff. Why over what.
+
+## Rules
+
+**Subject line:**
+- `<type>(<scope>): <imperative summary>` — `<scope>` optional
+- Types: `feat`, `fix`, `refactor`, `perf`, `docs`, `test`, `chore`, `build`, `ci`, `style`, `revert`
+- Imperative mood: "add", "fix", "remove" — not "added", "adds", "adding"
+- ≤50 chars when possible, hard cap 72
+- No trailing period
+- Match project convention for capitalization after the colon
+
+**Body (only if needed):**
+- Skip entirely when subject is self-explanatory
+- Add body only for: non-obvious *why*, breaking changes, migration notes, linked issues
+- Wrap at 72 chars
+- Bullets `-` not `*`
+- Reference issues/PRs at end: `Closes #42`, `Refs #17`
+
+**What NEVER goes in:**
+- "This commit does X", "I", "we", "now", "currently" — the diff says what
+- "As requested by..." — use Co-authored-by trailer
+- "Generated with Claude Code" or any AI attribution
+- Emoji (unless project convention requires)
+- Restating the file name when scope already says it
+
+## Examples
+
+Diff: new endpoint for user profile with body explaining the why
+- ❌ "feat: add a new endpoint to get user profile information from the database"
+- ✅
+  ```
+  feat(api): add GET /users/:id/profile
+
+  Mobile client needs profile data without the full user payload
+  to reduce LTE bandwidth on cold-launch screens.
+
+  Closes #128
+  ```
+
+Diff: breaking API change
+- ✅
+  ```
+  feat(api)!: rename /v1/orders to /v1/checkout
+
+  BREAKING CHANGE: clients on /v1/orders must migrate to /v1/checkout
+  before 2026-06-01. Old route returns 410 after that date.
+  ```
+
+## Auto-Clarity
+
+Always include body for: breaking changes, security fixes, data migrations, anything reverting a prior commit. Never compress these into subject-only — future debuggers need the context.
+
+## Boundaries
+
 Only generates the commit message. Does not run `git commit`, does not stage files, does not amend. Output the message as a code block ready to paste. "stop caveman-commit" or "normal mode": revert to verbose commit style.

imported-skills/julius-caveman/skills/caveman-compress/README.md

@@ -1,163 +1,163 @@
-<p align="center">
-  <img src="https://em-content.zobj.net/source/apple/391/rock_1faa8.png" width="80" />
-</p>
-
-<h1 align="center">caveman-compress</h1>
-
-<p align="center">
-  <strong>shrink memory file. save token every session.</strong>
-</p>
-
----
-
-A Claude Code skill that compresses your project memory files (`CLAUDE.md`, todos, preferences) into caveman format — so every session loads fewer tokens automatically.
-
-Claude read `CLAUDE.md` on every session start. If file big, cost big. Caveman make file small. Cost go down forever.
-
-## What It Do
-
-```
-/caveman-compress CLAUDE.md
-```
-
-```
-CLAUDE.md          ← compressed (Claude reads this — fewer tokens every session)
-CLAUDE.original.md ← human-readable backup (you edit this)
-```
-
-Original never lost. You can read and edit `.original.md`. Run skill again to re-compress after edits.
-
-## Benchmarks
-
-Real results on real project files:
-
-| File | Original | Compressed | Saved |
-|------|----------:|----------:|------:|
-| `claude-md-preferences.md` | 706 | 285 | **59.6%** |
-| `project-notes.md` | 1145 | 535 | **53.3%** |
-| `claude-md-project.md` | 1122 | 636 | **43.3%** |
-| `todo-list.md` | 627 | 388 | **38.1%** |
-| `mixed-with-code.md` | 888 | 560 | **36.9%** |
-| **Average** | **898** | **481** | **46%** |
-
-All validations passed ✅ — headings, code blocks, URLs, file paths preserved exactly.
-
-## Before / After
-
-<table>
-<tr>
-<td width="50%">
-
-### 📄 Original (706 tokens)
-
-> "I strongly prefer TypeScript with strict mode enabled for all new code. Please don't use `any` type unless there's genuinely no way around it, and if you do, leave a comment explaining the reasoning. I find that taking the time to properly type things catches a lot of bugs before they ever make it to runtime."
-
-</td>
-<td width="50%">
-
-### 🪨 Caveman (285 tokens)
-
-> "Prefer TypeScript strict mode always. No `any` unless unavoidable — comment why if used. Proper types catch bugs early."
-
-</td>
-</tr>
-</table>
-
-**Same instructions. 60% fewer tokens. Every. Single. Session.**
-
-## Security
-
-`caveman-compress` is flagged as Snyk High Risk due to subprocess and file I/O patterns detected by static analysis. This is a false positive — see [SECURITY.md](./SECURITY.md) for a full explanation of what the skill does and does not do.
-
-## Install
-
-Compress is built in with the `caveman` plugin. Install `caveman` once, then use `/caveman-compress`.
-
-If you need local files, the compress skill lives at:
-
-```bash
-caveman-compress/
-```
-
-**Requires:** Python 3.10+
-
-## Usage
-
-```
-/caveman-compress <filepath>
-```
-
-Examples:
-```
-/caveman-compress CLAUDE.md
-/caveman-compress docs/preferences.md
-/caveman-compress todos.md
-```
-
-### What files work
-
-| Type | Compress? |
-|------|-----------|
-| `.md`, `.txt`, `.rst`, `.typ`, `.typst`, `.tex` | ✅ Yes |
-| Extensionless natural language | ✅ Yes |
-| `.py`, `.js`, `.ts`, `.json`, `.yaml` | ❌ Skip (code/config) |
-| `*.original.md` | ❌ Skip (backup files) |
-
-## How It Work
-
-```
-/caveman-compress CLAUDE.md
-        ↓
-detect file type        (no tokens)
-        ↓
-Claude compresses       (tokens — one call)
-        ↓
-validate output         (no tokens)
-  checks: headings, code blocks, URLs, file paths, bullets
-        ↓
-if errors: Claude fixes cherry-picked issues only   (tokens — targeted fix)
-  does NOT recompress — only patches broken parts
-        ↓
-retry up to 2 times
-        ↓
-write compressed → CLAUDE.md
-write original   → CLAUDE.original.md
-```
-
-Only two things use tokens: initial compression + targeted fix if validation fails. Everything else is local Python.
-
-## What Is Preserved
-
-Caveman compress natural language. It never touch:
-
-- Code blocks (` ``` ` fenced or indented)
-- Inline code (`` `backtick content` ``)
-- URLs and links
-- File paths (`/src/components/...`)
-- Commands (`npm install`, `git commit`)
-- Technical terms, library names, API names
-- Headings (exact text preserved)
-- Tables (structure preserved, cell text compressed)
-- Dates, version numbers, numeric values
-
-## Why This Matter
-
-`CLAUDE.md` loads on **every session start**. A 1000-token project memory file costs tokens every single time you open a project. Over 100 sessions that's 100,000 tokens of overhead — just for context you already wrote.
-
-Caveman cut that by ~46% on average. Same instructions. Same accuracy. Less waste.
-
-```
-┌────────────────────────────────────────────┐
-│  TOKEN SAVINGS PER FILE    █████       46% │
-│  SESSIONS THAT BENEFIT     ██████████ 100% │
-│  INFORMATION PRESERVED     ██████████ 100% │
-│  SETUP TIME                █            1x │
-└─────────���──────────────────────────────────┘
-```
-
-## Part of Caveman
-
-This skill is part of the [caveman](https://github.com/JuliusBrussee/caveman) toolkit — making Claude use fewer tokens without losing accuracy.
-
-- **caveman** — make Claude *speak* like caveman (cuts response tokens ~65%)
-- **caveman-compress** — make Claude *read* less (cuts context tokens ~46%)
+<p align="center">
+  <img src="https://em-content.zobj.net/source/apple/391/rock_1faa8.png" width="80" />
+</p>
+
+<h1 align="center">caveman-compress</h1>
+
+<p align="center">
+  <strong>shrink memory file. save token every session.</strong>
+</p>
+
+---
+
+A Claude Code skill that compresses your project memory files (`CLAUDE.md`, todos, preferences) into caveman format — so every session loads fewer tokens automatically.
+
+Claude read `CLAUDE.md` on every session start. If file big, cost big. Caveman make file small. Cost go down forever.
+
+## What It Do
+
+```
+/caveman-compress CLAUDE.md
+```
+
+```
+CLAUDE.md          ← compressed (Claude reads this — fewer tokens every session)
+CLAUDE.original.md ← human-readable backup (you edit this)
+```
+
+Original never lost. You can read and edit `.original.md`. Run skill again to re-compress after edits.
+
+## Benchmarks
+
+Real results on real project files:
+
+| File | Original | Compressed | Saved |
+|------|----------:|----------:|------:|
+| `claude-md-preferences.md` | 706 | 285 | **59.6%** |
+| `project-notes.md` | 1145 | 535 | **53.3%** |
+| `claude-md-project.md` | 1122 | 636 | **43.3%** |
+| `todo-list.md` | 627 | 388 | **38.1%** |
+| `mixed-with-code.md` | 888 | 560 | **36.9%** |
+| **Average** | **898** | **481** | **46%** |
+
+All validations passed ✅ — headings, code blocks, URLs, file paths preserved exactly.
+
+## Before / After
+
+<table>
+<tr>
+<td width="50%">
+
+### 📄 Original (706 tokens)
+
+> "I strongly prefer TypeScript with strict mode enabled for all new code. Please don't use `any` type unless there's genuinely no way around it, and if you do, leave a comment explaining the reasoning. I find that taking the time to properly type things catches a lot of bugs before they ever make it to runtime."
+
+</td>
+<td width="50%">
+
+### 🪨 Caveman (285 tokens)
+
+> "Prefer TypeScript strict mode always. No `any` unless unavoidable — comment why if used. Proper types catch bugs early."
+
+</td>
+</tr>
+</table>
+
+**Same instructions. 60% fewer tokens. Every. Single. Session.**
+
+## Security
+
+`caveman-compress` is flagged as Snyk High Risk due to subprocess and file I/O patterns detected by static analysis. This is a false positive — see [SECURITY.md](./SECURITY.md) for a full explanation of what the skill does and does not do.
+
+## Install
+
+Compress is built in with the `caveman` plugin. Install `caveman` once, then use `/caveman-compress`.
+
+If you need local files, the compress skill lives at:
+
+```bash
+caveman-compress/
+```
+
+**Requires:** Python 3.10+
+
+## Usage
+
+```
+/caveman-compress <filepath>
+```
+
+Examples:
+```
+/caveman-compress CLAUDE.md
+/caveman-compress docs/preferences.md
+/caveman-compress todos.md
+```
+
+### What files work
+
+| Type | Compress? |
+|------|-----------|
+| `.md`, `.txt`, `.rst`, `.typ`, `.typst`, `.tex` | ✅ Yes |
+| Extensionless natural language | ✅ Yes |
+| `.py`, `.js`, `.ts`, `.json`, `.yaml` | ❌ Skip (code/config) |
+| `*.original.md` | ❌ Skip (backup files) |
+
+## How It Work
+
+```
+/caveman-compress CLAUDE.md
+        ↓
+detect file type        (no tokens)
+        ↓
+Claude compresses       (tokens — one call)
+        ↓
+validate output         (no tokens)
+  checks: headings, code blocks, URLs, file paths, bullets
+        ↓
+if errors: Claude fixes cherry-picked issues only   (tokens — targeted fix)
+  does NOT recompress — only patches broken parts
+        ↓
+retry up to 2 times
+        ↓
+write compressed → CLAUDE.md
+write original   → CLAUDE.original.md
+```
+
+Only two things use tokens: initial compression + targeted fix if validation fails. Everything else is local Python.
+
+## What Is Preserved
+
+Caveman compress natural language. It never touch:
+
+- Code blocks (` ``` ` fenced or indented)
+- Inline code (`` `backtick content` ``)
+- URLs and links
+- File paths (`/src/components/...`)
+- Commands (`npm install`, `git commit`)
+- Technical terms, library names, API names
+- Headings (exact text preserved)
+- Tables (structure preserved, cell text compressed)
+- Dates, version numbers, numeric values
+
+## Why This Matter
+
+`CLAUDE.md` loads on **every session start**. A 1000-token project memory file costs tokens every single time you open a project. Over 100 sessions that's 100,000 tokens of overhead — just for context you already wrote.
+
+Caveman cut that by ~46% on average. Same instructions. Same accuracy. Less waste.
+
+```
+┌────────────────────────────────────────────┐
+│  TOKEN SAVINGS PER FILE    █████       46% │
+│  SESSIONS THAT BENEFIT     ██████████ 100% │
+│  INFORMATION PRESERVED     ██████████ 100% │
+│  SETUP TIME                █            1x │
+└────────────────────────────────────────────┘
+```
+
+## Part of Caveman
+
+This skill is part of the [caveman](https://github.com/JuliusBrussee/caveman) toolkit — making Claude use fewer tokens without losing accuracy.
+
+- **caveman** — make Claude *speak* like caveman (cuts response tokens ~65%)
+- **caveman-compress** — make Claude *read* less (cuts context tokens ~46%)

imported-skills/julius-caveman/skills/caveman-compress/SECURITY.md

@@ -1,31 +1,31 @@
-# Security
-
-## Snyk High Risk Rating
-
-`caveman-compress` receives a Snyk High Risk rating due to static analysis heuristics. This document explains what the skill does and does not do.
-
-### What triggers the rating
-
-1. **subprocess usage**: The skill calls the `claude` CLI via `subprocess.run()` as a fallback when `ANTHROPIC_API_KEY` is not set. The subprocess call uses a fixed argument list — no shell interpolation occurs. User file content is passed via stdin, not as a shell argument.
-
-2. **File read/write**: The skill reads the file the user explicitly points it at, compresses it, and writes the result back to the same path. A `.original.md` backup is saved alongside it. No files outside the user-specified path are read or written.
-
-### What the skill does NOT do
-
-- Does not execute user file content as code
-- Does not make network requests except to Anthropic's API (via SDK or CLI)
-- Does not access files outside the path the user provides
-- Does not use shell=True or string interpolation in subprocess calls
-- Does not collect or transmit any data beyond the file being compressed
-
-### Auth behavior
-
-If `ANTHROPIC_API_KEY` is set, the skill uses the Anthropic Python SDK directly (no subprocess). If not set, it falls back to the `claude` CLI, which uses the user's existing Claude desktop authentication.
-
-### File size limit
-
-Files larger than 500KB are rejected before any API call is made.
-
-### Reporting a vulnerability
-
-If you believe you've found a genuine security issue, please open a GitHub issue with the label `security`.
+# Security
+
+## Snyk High Risk Rating
+
+`caveman-compress` receives a Snyk High Risk rating due to static analysis heuristics. This document explains what the skill does and does not do.
+
+### What triggers the rating
+
+1. **subprocess usage**: The skill calls the `claude` CLI via `subprocess.run()` as a fallback when `ANTHROPIC_API_KEY` is not set. The subprocess call uses a fixed argument list — no shell interpolation occurs. User file content is passed via stdin, not as a shell argument.
+
+2. **File read/write**: The skill reads the file the user explicitly points it at, compresses it, and writes the result back to the same path. A `.original.md` backup is saved alongside it. No files outside the user-specified path are read or written.
+
+### What the skill does NOT do
+
+- Does not execute user file content as code
+- Does not make network requests except to Anthropic's API (via SDK or CLI)
+- Does not access files outside the path the user provides
+- Does not use shell=True or string interpolation in subprocess calls
+- Does not collect or transmit any data beyond the file being compressed
+
+### Auth behavior
+
+If `ANTHROPIC_API_KEY` is set, the skill uses the Anthropic Python SDK directly (no subprocess). If not set, it falls back to the `claude` CLI, which uses the user's existing Claude desktop authentication.
+
+### File size limit
+
+Files larger than 500KB are rejected before any API call is made.
+
+### Reporting a vulnerability
+
+If you believe you've found a genuine security issue, please open a GitHub issue with the label `security`.

imported-skills/julius-caveman/skills/caveman-compress/SKILL.md

@@ -1,111 +1,111 @@
----
-name: caveman-compress
-description: >
-  Compress natural language memory files (CLAUDE.md, todos, preferences) into caveman format
-  to save input tokens. Preserves all technical substance, code, URLs, and structure.
-  Compressed version overwrites the original file. Human-readable backup saved as FILE.original.md.
-  Trigger: /caveman-compress FILEPATH or "compress memory file"
----
-
-# Caveman Compress
-
-## Purpose
-
-Compress natural language files (CLAUDE.md, todos, preferences) into caveman-speak to reduce input tokens. Compressed version overwrites original. Human-readable backup saved as `<filename>.original.md`.
-
-## Trigger
-
-`/caveman-compress <filepath>` or when user asks to compress a memory file.
-
-## Process
-
-1. The compression scripts live in `scripts/` (adjacent to this SKILL.md). If the path is not immediately available, search for `scripts/__main__.py` next to this SKILL.md.
-
-2. From the directory containing this SKILL.md, run:
-
-python3 -m scripts <absolute_filepath>
-
-3. The CLI will:
-- detect file type (no tokens)
-- call Claude to compress
-- validate output (no tokens)
-- if errors: cherry-pick fix with Claude (targeted fixes only, no recompression)
-- retry up to 2 times
-- if still failing after 2 retries: report error to user, leave original file untouched
-
-4. Return result to user
-
-## Compression Rules
-
-### Remove
-- Articles: a, an, the
-- Filler: just, really, basically, actually, simply, essentially, generally
-- Pleasantries: "sure", "certainly", "of course", "happy to", "I'd recommend"
-- Hedging: "it might be worth", "you could consider", "it would be good to"
-- Redundant phrasing: "in order to" → "to", "make sure to" → "ensure", "the reason is because" → "because"
-- Connective fluff: "however", "furthermore", "additionally", "in addition"
-
-### Preserve EXACTLY (never modify)
-- Code blocks (fenced ``` and indented)
-- Inline code (`backtick content`)
-- URLs and links (full URLs, markdown links)
-- File paths (`/src/components/...`, `./config.yaml`)
-- Commands (`npm install`, `git commit`, `docker build`)
-- Technical terms (library names, API names, protocols, algorithms)
-- Proper nouns (project names, people, companies)
-- Dates, version numbers, numeric values
-- Environment variables (`$HOME`, `NODE_ENV`)
-
-### Preserve Structure
-- All markdown headings (keep exact heading text, compress body below)
-- Bullet point hierarchy (keep nesting level)
-- Numbered lists (keep numbering)
-- Tables (compress cell text, keep structure)
-- Frontmatter/YAML headers in markdown files
-
-### Compress
-- Use short synonyms: "big" not "extensive", "fix" not "implement a solution for", "use" not "utilize"
-- Fragments OK: "Run tests before commit" not "You should always run tests before committing"
-- Drop "you should", "make sure to", "remember to" — just state the action
-- Merge redundant bullets that say the same thing differently
-- Keep one example where multiple examples show the same pattern
-
-CRITICAL RULE:
-Anything inside ``` ... ``` must be copied EXACTLY.
-Do not:
-- remove comments
-- remove spacing
-- reorder lines
-- shorten commands
-- simplify anything
-
-Inline code (`...`) must be preserved EXACTLY.
-Do not modify anything inside backticks.
-
-If file contains code blocks:
-- Treat code blocks as read-only regions
-- Only compress text outside them
-- Do not merge sections around code
-
-## Pattern
-
-Original:
-> You should always make sure to run the test suite before pushing any changes to the main branch. This is important because it helps catch bugs early and prevents broken builds from being deployed to production.
-
-Compressed:
-> Run tests before push to main. Catch bugs early, prevent broken prod deploys.
-
-Original:
-> The application uses a microservices architecture with the following components. The API gateway handles all incoming requests and routes them to the appropriate service. The authentication service is responsible for managing user sessions and JWT tokens.
-
-Compressed:
-> Microservices architecture. API gateway route all requests to services. Auth service manage user sessions + JWT tokens.
-
-## Boundaries
-
-- ONLY compress natural language files (.md, .txt, .typ, .typst, .tex, extensionless)
-- NEVER modify: .py, .js, .ts, .json, .yaml, .yml, .toml, .env, .lock, .css, .html, .xml, .sql, .sh
-- If file has mixed content (prose + code), compress ONLY the prose sections
-- If unsure whether something is code or prose, leave it unchanged
-- Original file is backed up as FILE.original.md before overwriting
-- Never compress FILE.original.md (skip it)
+---
+name: caveman-compress
+description: >
+  Compress natural language memory files (CLAUDE.md, todos, preferences) into caveman format
+  to save input tokens. Preserves all technical substance, code, URLs, and structure.
+  Compressed version overwrites the original file. Human-readable backup saved as FILE.original.md.
+  Trigger: /caveman-compress FILEPATH or "compress memory file"
+---
+
+# Caveman Compress
+
+## Purpose
+
+Compress natural language files (CLAUDE.md, todos, preferences) into caveman-speak to reduce input tokens. Compressed version overwrites original. Human-readable backup saved as `<filename>.original.md`.
+
+## Trigger
+
+`/caveman-compress <filepath>` or when user asks to compress a memory file.
+
+## Process
+
+1. The compression scripts live in `scripts/` (adjacent to this SKILL.md). If the path is not immediately available, search for `scripts/__main__.py` next to this SKILL.md.
+
+2. From the directory containing this SKILL.md, run:
+
+python3 -m scripts <absolute_filepath>
+
+3. The CLI will:
+- detect file type (no tokens)
+- call Claude to compress
+- validate output (no tokens)
+- if errors: cherry-pick fix with Claude (targeted fixes only, no recompression)
+- retry up to 2 times
+- if still failing after 2 retries: report error to user, leave original file untouched
+
+4. Return result to user
+
+## Compression Rules
+
+### Remove
+- Articles: a, an, the
+- Filler: just, really, basically, actually, simply, essentially, generally
+- Pleasantries: "sure", "certainly", "of course", "happy to", "I'd recommend"
+- Hedging: "it might be worth", "you could consider", "it would be good to"
+- Redundant phrasing: "in order to" → "to", "make sure to" → "ensure", "the reason is because" → "because"
+- Connective fluff: "however", "furthermore", "additionally", "in addition"
+
+### Preserve EXACTLY (never modify)
+- Code blocks (fenced ``` and indented)
+- Inline code (`backtick content`)
+- URLs and links (full URLs, markdown links)
+- File paths (`/src/components/...`, `./config.yaml`)
+- Commands (`npm install`, `git commit`, `docker build`)
+- Technical terms (library names, API names, protocols, algorithms)
+- Proper nouns (project names, people, companies)
+- Dates, version numbers, numeric values
+- Environment variables (`$HOME`, `NODE_ENV`)
+
+### Preserve Structure
+- All markdown headings (keep exact heading text, compress body below)
+- Bullet point hierarchy (keep nesting level)
+- Numbered lists (keep numbering)
+- Tables (compress cell text, keep structure)
+- Frontmatter/YAML headers in markdown files
+
+### Compress
+- Use short synonyms: "big" not "extensive", "fix" not "implement a solution for", "use" not "utilize"
+- Fragments OK: "Run tests before commit" not "You should always run tests before committing"
+- Drop "you should", "make sure to", "remember to" — just state the action
+- Merge redundant bullets that say the same thing differently
+- Keep one example where multiple examples show the same pattern
+
+CRITICAL RULE:
+Anything inside ``` ... ``` must be copied EXACTLY.
+Do not:
+- remove comments
+- remove spacing
+- reorder lines
+- shorten commands
+- simplify anything
+
+Inline code (`...`) must be preserved EXACTLY.
+Do not modify anything inside backticks.
+
+If file contains code blocks:
+- Treat code blocks as read-only regions
+- Only compress text outside them
+- Do not merge sections around code
+
+## Pattern
+
+Original:
+> You should always make sure to run the test suite before pushing any changes to the main branch. This is important because it helps catch bugs early and prevents broken builds from being deployed to production.
+
+Compressed:
+> Run tests before push to main. Catch bugs early, prevent broken prod deploys.
+
+Original:
+> The application uses a microservices architecture with the following components. The API gateway handles all incoming requests and routes them to the appropriate service. The authentication service is responsible for managing user sessions and JWT tokens.
+
+Compressed:
+> Microservices architecture. API gateway route all requests to services. Auth service manage user sessions + JWT tokens.
+
+## Boundaries
+
+- ONLY compress natural language files (.md, .txt, .typ, .typst, .tex, extensionless)
+- NEVER modify: .py, .js, .ts, .json, .yaml, .yml, .toml, .env, .lock, .css, .html, .xml, .sql, .sh
+- If file has mixed content (prose + code), compress ONLY the prose sections
+- If unsure whether something is code or prose, leave it unchanged
+- Original file is backed up as FILE.original.md before overwriting
+- Never compress FILE.original.md (skip it)