docs/dev/open_tasks.md

Open Tasks

Single source of truth for active work.

Priority Legend

• P0 = blocking production/demo
• P1 = high impact
• P2 = nice to have

Active

ID	Priority	Task	Owner	Status	Next Step	Verification
T-137	P2	Keep repo hygiene bounded by pruning non-canonical eval artifacts while preserving canonical eval baselines and runner assets	Engineering	DONE	Reuse backend/scripts/cleanup_repo_artifacts.py --dry-run --no-backup periodically instead of letting timestamped eval reruns accumulate in git	python3 backend/scripts/cleanup_repo_artifacts.py --no-backup -> removed_dirs=75, removed_files=62; follow-up python3 backend/scripts/cleanup_repo_artifacts.py --dry-run --no-backup -> 0 pending removals; canonical probes for latest_eval25_guarded_gpt_check, latest_eval50_guarded_gpt_check, latest_eval6_concept_check, release_gate, shards10, and shards5_eval75 all returned OK
T-146	P1	Add a real Hugging Face canary target so the gated deploy workflow exercises both canary and production lanes instead of production-only	Engineering	DONE	Keep the canary Space config aligned with production except for host-specific APP_BASE_URL / VITE_APP_BASE_URL, and keep HF_SPACE_ID_CANARY populated in GitHub Actions	GitHub Actions secret HF_SPACE_ID_CANARY now points at crazycrazypete/Masters-four-Tab-OpenAI-Canary; Actions run 22813479490 finished with both deploy-canary and deploy-production green after the canary host-specific base URL was corrected
T-147	P1	Run the first authenticated hosted smoke pass against the refreshed production build and use that result to close the stale hosted sign-off loop	Engineering	DONE	Reuse the same minimal hosted smoke set after future deploys: auth full-flow, one assistant-family provider query, and one POTS workspace shell check on both canary and production when relevant	Credentialed hosted smoke passed on both production and canary: cd frontend && npx playwright test e2e/auth.full-flow.spec.ts --reporter=line -> 1 passed; cd frontend && E2E_BASE_URL=https://crazycrazypete-masters-four-tab-openai-canary.hf.space npx playwright test e2e/auth.full-flow.spec.ts --reporter=line -> 1 passed; cd frontend && npx playwright test e2e/pots.provider-coverage.spec.ts --reporter=line -> 1 passed; cd frontend && E2E_BASE_URL=https://crazycrazypete-masters-four-tab-openai-canary.hf.space npx playwright test e2e/pots.provider-coverage.spec.ts --reporter=line -> 1 passed; one-off headless POTS workspace smoke passed on both hosts and confirmed the POTS Project Workspace shell is live instead of the old stacked page
D-232	Removed the duplicate assistant-tab security/CAPTCHA checks from the shared Help + Assist launcher, Unified Knowledgebase, and POTS assistant flows while keeping the Rapid Router order-submit CAPTCHA intact	2026-03-07	backend/app/main.py, backend/app/test_knowledgebase_api.py, backend/app/test_chat_guidance_api.py, frontend/src/components/FloatingRouterHelper.tsx, frontend/src/pages/UnifiedKnowledgebase.tsx, frontend/src/pages/PotsAssistant.tsx, frontend/src/pages/RapidRouter.tsx; cd backend && .venv/bin/python -m pytest -q app/test_knowledgebase_api.py app/test_chat_guidance_api.py app/test_rapid_router_api_shell.py -> 37 passed; cd frontend && npm run build -> success; helper Vitest file remains blocked by local worker hang after startup
T-144	P1	Remove duplicate per-tab assistant security checks from the shared help/assistant tabs while preserving the Rapid Router order-submit CAPTCHA	Engineering	DONE	If assistant abuse appears later, add rate-limit/abuse controls at the backend instead of restoring per-tab CAPTCHA friction	cd backend && .venv/bin/python -m pytest -q app/test_knowledgebase_api.py app/test_chat_guidance_api.py app/test_rapid_router_api_shell.py -> 37 passed; cd frontend && npm run build -> success; cd frontend && npx vitest run src/components/FloatingRouterHelper.test.tsx --reporter=dot still stalls after startup in the current local Vitest worker environment
T-143	P1	Make Rapid Router advanced configuration notes optional whenever at least one advanced checkbox option is selected, while keeping notes required for freeform advanced requests with no selected task	Engineering	DONE	Keep backend/frontend validation aligned if new advanced task checkboxes are added later	cd backend && .venv/bin/python -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py -> 53 passed; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 6 passed; cd frontend && npm run build -> success
T-142	P1	Add the four new required Rapid Router approval attestations and enforce them server-side in the order submit path	Engineering	DONE	Keep any future approval-copy changes mirrored in both frontend validation and backend approvals schema so order-submit rules cannot drift	cd backend && .venv/bin/python -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py -> 53 passed; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 5 passed; cd frontend && npm run build -> success
D-231	Changed Rapid Router BoBo bill-to phone to a full 10-digit US phone field with (111) 222-2222 formatting, matching frontend validation, backend normalization, and rendered PDF/email output	2026-03-07	frontend/src/pages/RapidRouter.tsx, frontend/src/pages/RapidRouter.test.tsx, backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py; cd backend && .venv/bin/python -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py -> 52 passed; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 5 passed; cd frontend && npm run build -> success
T-141	P1	Align Rapid Router BoBo bill-to phone UX and validation with the requested full-phone example (111) 222-2222 instead of the legacy 7-digit local-number rule	Engineering	DONE	Keep BoBo bill-to phone on the same full 10-digit validation/rendering path as other US phone fields unless the business later confirms a local-only requirement	cd backend && .venv/bin/python -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py -> 52 passed; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 5 passed; cd frontend && npm run build -> success
D-230	Rapid Router split shipping now clamps per-location assignment to ordered quantity, disables adding more locations once all units are assigned, and persists the optional Configure IP passthrough advanced task through backend normalization and output rendering	2026-03-07	frontend/src/pages/RapidRouter.tsx, frontend/src/pages/RapidRouter.test.tsx, backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py; cd backend && .venv/bin/python -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py -> 52 passed; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 5 passed; cd frontend && npm run build -> success
T-140	P1	Keep Rapid Router split-shipping allocations bounded to total cart quantity and add the optional advanced Configure IP passthrough task through submit/output paths	Engineering	DONE	If shipping allocation rules expand later, keep clamping at edit-time instead of allowing invalid temporary over-assignment states	cd backend && .venv/bin/python -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py -> 52 passed; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 5 passed; cd frontend && npm run build -> success
T-138	P1	Close the last local continuity gap on the former guarded-GPT Masters mention-lookup tail (31, 32, 35, 37) and then validate that improvement in the next broader rerun	Engineering	DONE	Keep _should_skip_masters_concept_preflight(...) aligned with the explicit doc-lookup vocabulary so plural forms like documents, docs, files, and sources never reactivate concept fallback for file-title lookups	Focused slice stayed fast: python3 backend/scripts/unified_kb_eval150.py --cases /tmp/mtk_focused_eval_slices/masters_slice_cases.json ... -> 4/4 passed, avg 7.13ms, p95 26.42ms; exact .env.codex repros dropped from multi-second latency to fast-path timings (31: ~`2461ms->28.83ms, 32: ~2635ms->26.95ms); broader 31-40rerun ->10/10 passed, avg 5.12ms, Masters avg 4.19ms, p95 26.91ms`
T-139	P1	Finish trimming the remaining reusable router compare/render tail so the broader guarded suites can be rerun against a materially cleaner router latency baseline	Engineering	DONE	Keep the shared-query compare/antenna pattern in place unless a future rerun shows router regressions again; router is no longer the first broad-suite blocker	Focused slice after the shared-query pass: python3 backend/scripts/unified_kb_eval150.py --cases /tmp/mtk_focused_eval_slices/router_slice_cases.json ... -> 7/7 passed, avg 145.25ms, router-doc avg 328.76ms, p95 627.21ms; broader 75 rerun in docs/evals/20260308_guarded75_after_masters_fix/ kept router compare prompts materially lower (42 ~`327ms, 114 ~661ms) while overall suite finished 75/75, avg_latency_ms=28.81, p95_ms=55.15`
T-145	P1	Re-run the focused Rapid Router frontend validation chain once the local Codex unified-exec saturation is cleared so the new browse-first + BoBo/authorization flow has a clean frontend automated verification summary	Engineering	DONE	Keep future focused frontend reruns isolated so local exec saturation does not masquerade as product regressions	cd backend && .venv/bin/python -m pytest -x -vv app/rapid_router/test_rapid_router_core.py -> 28 passed; cd backend && .venv/bin/python -m pytest -q app/test_rapid_router_api_shell.py -> 24 passed; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 3 passed; cd frontend && npm run build -> success
T-137	P1	Eliminate the remaining guarded-GPT delegate latency tails now that the former Masters mention bucket and the main router compare bucket are no longer the broad-suite blocker	Engineering	IN PROGRESS	Chase the new dominant tails next: POTS playbook/provider prompts (79/82/86) first, then the smaller longer-form Masters content-pack cluster (97/99/101, 106/111/134) plus the lingering BuSS-SKU doc prompt (30)	docs/evals/20260308_guarded75_after_masters_fix/unified_kb_eval150_shards10_summary.json -> 75/75, avg_latency_ms=28.81, p95_ms=55.15, p99_ms=327.53, stage_budget_exits=0; docs/evals/20260308_guarded150_after_masters_fix/unified_kb_eval150_shards10_summary.json -> 150/150, avg_latency_ms=151.14, p95_ms=661.36, p99_ms=2969.52, stage_budget_exits=0; remaining dominant outliers are 79/82/86 at 2.88s-4.36s, with secondary longer-form Masters prompts 97/99/101, 106/111/134 and case 30 (725ms)
T-133	P1	Trim the residual deterministic/content-assembly latency that remains after the Masters mention and router compare fixes	Engineering	OPEN	Start with the old POTS playbook/provider bucket (79/82/86), then profile the longer-form Masters content-pack prompts before paying for another promotion-quality baseline rerun	docs/evals/20260308_guarded150_after_masters_fix/unified_kb_eval150_shards10_summary.json -> 150/150, p95_ms=661.36, p99_ms=2969.52; dominant remaining tails: 79 2879.88ms, 82 2969.52ms, 86 4359.05ms; secondary longer-form Masters/content prompts: 97 334.73ms, 99 639.50ms, 101 661.69ms, 106 669.12ms, 111 315.54ms, 134 773.01ms
T-132	P1	Rerun broader guarded-GPT 75 and 150 evals against the new deterministic fast answers, exact/current blocked-case net, and the now-green 50-case concept pack	Engineering	DONE	Use the rerun artifacts as comparison evidence, but keep 25/50 as the stable lightweight gates until the new 75/150 regressions in T-133 are fixed	docs/evals/20260307_010031_eval75_guarded_gpt_rerun/unified_kb_eval150_shards10_summary.json -> 75 / 75 passed; docs/evals/20260307_010031_eval150_guarded_gpt_rerun/unified_kb_eval150_shards10_summary.json -> 149 / 150 passed; bash backend/scripts/test_backend.sh --full remained previously green at 501 passed
T-131	P1	Confirm hosted deployment env does not override the new gpt-5-mini repo defaults with stale OPENAI_MODEL or assistant-specific model pins	Engineering	OPEN	Inspect Hugging Face Space secrets/variables and any production deployment env to ensure OPENAI_MODEL, UNIFIED_KB_OPENAI_MODEL, and ROUTER_RAG_OPENAI_MODEL are unset or explicitly gpt-5-mini, then rerun one live assistant smoke check	Local repo/runtime verification is complete, but hosted env values are external to git and can still override repo defaults
T-130	P1	Standardize all active LLM-assisted backend/runtime defaults, env examples, and local repo env pins on gpt-5-mini	Engineering	DONE	Treat gpt-5-mini as the current default baseline and only revisit if a future model migration is deliberate and fully tested across backend + eval runners	rg -n 'gpt-5\\.2' README.md backend/app backend/scripts backend/.env.test.example .env.codex backend/.env.codex ... -> no active runtime/config hits; cd backend && .venv/bin/python -m pytest -q app/test_pots_conversation_regression.py -k 'concept_fallback_for_generic_pots_question or llm_synthesis_omits_temperature_for_gpt5_models' -> 2 passed; bash backend/scripts/test_backend.sh --full -> 478 passed; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npm run build -> success; cd frontend && npm run test -> 31 files / 111 passed; docs/evals/20260306_230403_eval25_gpt5mini_default/unified_kb_eval150_shards10_summary.json -> 25 / 25 passed
T-129	P2	Stabilize the residual guarded-GPT dual-pathway explainer and rerun the reusable 25-case pack	Engineering	DONE	Keep the dual-pathway phrasing under regression watch as future concept-fallback work lands; no additional action is required for the current acceptance gate	docs/evals/20260306_230403_eval25_gpt5mini_default/unified_kb_eval150_shards10_summary.json -> 25 / 25 passed, failed_ids=[]
T-128	P1	Create a reusable guarded-GPT acceptance pack with 25 questions split into 5-question shards and a dedicated shard runner	Engineering	DONE	Use this pack as the lightweight regression gate for future guarded-GPT changes; only expand it after stabilizing or replacing residual case 13	python3 - <<'PY' ... len(rows)==25 ... PY -> success; bash -n backend/scripts/run_unified_kb_eval25_guarded_gpt_chunks.sh -> success; docs/evals/20260307_001201_eval25_phase12/unified_kb_eval150_shards10_summary.json -> 25 / 25 passed (100.0%)
T-127	P1	Roll out the shared assistant-family concept fallback chain with allow/deny gates, gpt-5-mini, provenance labels, and GPT+web only when the model-only concept answer still needs refinement	Engineering	DONE	Expand deterministic internal concept fast answers for the highest-frequency telecom/router/POTS explainers so the new fallback is used less often for questions that can be answered cheaply from curated internal patterns	cd backend && .venv/bin/python -m pytest -q app/test_assistant_fallback.py app/test_unified_kb_core.py app/test_router_rag_module.py app/test_masters_conversation_regression.py app/test_pots_conversation_regression.py app/test_chat_guidance_api.py app/test_knowledgebase_api.py -> 202 passed; bash backend/scripts/test_backend.sh --full -> 477 passed; Router RAG smoke -> 10 passed; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npm run build -> success; cd frontend && npm run test -> 31 files / 111 passed; docs/evals/latest_eval6_concept_check/unified_kb_eval150_shards10_summary.json -> 6 / 6 passed
T-126	P1	Redeploy the hosted Hugging Face app and rerun the live POTS provider-coverage Playwright spec so the local MetTel provider-card backfill is validated against the actual live site	Engineering	OPEN	Ship the current backend provider-card patch, wait for Hugging Face to rebuild, then rerun cd frontend && npx playwright test e2e/pots.provider-coverage.spec.ts --config=playwright.config.ts against the hosted base URL	Local fix coverage is green: cd backend && .venv/bin/python -m pytest -q app/test_unified_kb_core.py -k 'provider_inventory_supplements_missing_pots_provider_cards_from_router_corpus or provider_inventory_backfills_missing_router_hint_paths_from_index_hits' -> 2 passed; cd backend && .venv/bin/python -m pytest -q app/test_pots_provider_recall.py -> 2 passed; live suite remains 9 passed / 1 failed / 4 skipped until redeploy
T-125	P1	Enforce the current UI-lock scan rules by removing dead collapsed banners, hiding default status/debug entry points, and eliminating duplicate primary CTAs where they still leak into the active viewport	Engineering	DONE	Keep future cleanup focused on dense admin-only surfaces and message-detail consistency, not reopening already-compliant shell patterns	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/AssistantWorkspace.test.tsx src/components/PromptCoach.test.tsx src/components/BrandHeader.test.tsx src/pages/RapidRouter.test.tsx --reporter=dot -> 11 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 30 files / 105 passed; git diff --check -> success
T-124	P1	Lock the knowledge/chat family to one shared assistant shell, auto-collapse setup after the first user turn, and restyle the legacy assistant pages onto that pattern	Engineering	DONE	If assistant cleanup continues, unify the deeper response-detail treatments (Why, Next action, Sources, file panels) so all assistant answers share one internal message pattern as well	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/AssistantWorkspace.test.tsx src/components/PageArchetypes.test.tsx --reporter=dot -> 4 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 30 files / 105 passed
T-123	P1	Rebuild RapidRouter as a staged scan-and-build commerce flow (Filter, Browse, Quantity, Customer info, Review) with a sticky cart and collapsed secondary tools	Engineering	DONE	Collapse any remaining late-stage advanced/admin helper clusters behind one secondary control so the new commerce sequence stays clean under heavy use	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 2 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 29 files / 103 passed
T-122	P1	Collapse Telco assumptions, what-if mode, diagnostics, quote helpers, scenario JSON/CSV, and assistant coaching into one shared Advanced drawer so the default calculator surface stays on the business flow	Engineering	DONE	Apply the same single-secondary-control rule to RapidRouter, which still exposes too many business and support surfaces in parallel	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/TelcoCalculator.test.tsx --reporter=dot -> 2 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 28 files / 101 passed
T-121	P1	Rebuild TelcoCalculator as a single-path step sequence with Locations, Pricing, Results, and Export instead of a simultaneous tri-column calculator surface	Engineering	DONE	Apply the same step-led simplification to RapidRouter, which still mixes catalog, helper, and order-prep surfaces in one view	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/TelcoCalculator.test.tsx --reporter=dot -> 1 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 28 files / 100 passed
T-120	P1	Replace paragraph-style POTS instructions with a stable three-line step guide so each step only says what it does, what is needed now, and what happens next	Engineering	DONE	Reuse the same guide pattern in the POTS project drawer and any later summary/export surfaces that still read as prose-heavy	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsEstimateIntake.test.tsx src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsIntake.test.tsx src/pages/PotsWorkspace.test.tsx --reporter=dot -> 23 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 99 passed
T-119	P1	Flatten the embedded PotsEstimateIntake wrapper so the estimator and intake inherit a lighter host shell instead of stacking full cards inside cards	Engineering	DONE	If more simplification is still needed, flatten the later review/export sections inside PotsIntake rather than adding more wrapper-level chrome	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsEstimateIntake.test.tsx src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsIntake.test.tsx src/pages/PotsWorkspace.test.tsx --reporter=dot -> 23 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 99 passed
T-118	P1	Convert PotsWorkspace routing questions into a one-question-at-a-time conversation with answer cards and compact Why this matters disclosure	Engineering	DONE	Reuse the same guided-question pattern in other dense decision forms if later UI lock passes show similar cognitive overload	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsWorkspace.test.tsx --reporter=dot -> 10 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 99 passed
T-117	P1	Move active-project creation and saved-project switching behind the Project tools drawer so PotsWorkspace stops showing setup UI in the main wizard by default	Engineering	DONE	Apply the same hide-setup-behind-drawer rule in other dense workflows such as RapidRouter and TelcoCalculator where setup/admin surfaces still compete with the active task	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsWorkspace.test.tsx --reporter=dot -> 9 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 98 passed
T-116	P1	Convert PotsWorkspace into a true wizard shell with one active step and one optional utility drawer	Engineering	DONE	Apply the same step-led shell discipline to RapidRouter and TelcoCalculator if the UI lock continues beyond POTS	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsWorkspace.test.tsx --reporter=dot -> 8 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 97 passed; git diff --check -> success
T-115	P1	Enforce one obvious primary action per screen so setup, reset, export, and support controls stop competing with the current forward move	Engineering	DONE	Continue the CTA-hierarchy pass in RapidRouter, TelcoCalculator, and the assistant-family export/help clusters where multiple strong actions still share one viewport	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsEstimateIntake.test.tsx src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsWorkspace.test.tsx --reporter=dot -> 16 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 97 passed; git diff --check -> success
T-114	P1	Tighten and standardize the radius system so major shells use 20px, secondary surfaces use 16px, controls use 12px, and full-pill styling is reserved for chips	Engineering	DONE	Continue the same radius cleanup through RapidRouter, TelcoCalculator, CommandPalette, and any remaining legacy modal/helper surfaces that still overuse rounded-2xl	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/PrimaryNavigation.test.tsx src/components/FloatingRouterHelper.test.tsx src/components/PromptCoach.test.tsx src/components/chat/ChatTranscript.test.tsx src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsEstimateIntake.test.tsx src/pages/PotsIntake.test.tsx src/pages/PotsWorkspace.test.tsx --reporter=dot -> 34 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 97 passed; git diff --check -> success
T-113	P1	Replace border-heavy card stacking with the locked three-surface whitespace hierarchy in the shared shell and active POTS flow	Engineering	DONE	Continue the same whitespace-hierarchy cleanup in TelcoCalculator, RapidRouter, and the still-denser late-step surfaces in PotsIntake if the UI lock continues	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/PageArchetypes.test.tsx src/pages/PotsWorkspace.test.tsx src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsEstimateIntake.test.tsx src/pages/PotsIntake.test.tsx --reporter=dot -> 23 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 97 passed; git diff --check -> success
T-112	P1	Reduce badge and label noise in the shared shell, POTS flow, and assistant-family pages so metadata stops competing with primary actions	Engineering	DONE	If the badge-noise pass continues, target the remaining denser local-state surfaces next: TelcoCalculator, RapidRouter, and the still-busy parts of PotsIntake	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/FloatingRouterHelper.test.tsx src/components/PageArchetypes.test.tsx src/pages/PotsWorkspace.test.tsx src/pages/PotsSavingsEstimator.test.tsx --reporter=dot -> 19 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 97 passed
T-111	P1	Lock the shared typography system so the shell uses Public Sans, a slightly larger reading scale, and uppercase only for true metadata	Engineering	DONE	Continue the UI lock by applying the same shared typography utilities opportunistically to any remaining dense admin/reporting surfaces as later layout passes touch them	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/BrandHeader.test.tsx src/components/PrimaryNavigation.test.tsx src/components/PageArchetypes.test.tsx src/components/FloatingRouterHelper.test.tsx src/pages/PotsWorkspace.test.tsx src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsEstimateIntake.test.tsx src/pages/PotsIntake.test.tsx --reporter=dot -> 36 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 97 passed; typography-scope grep confirmed no remaining uppercase classes in the shared shell + active assistant/POTS lock scope
T-110	P1	Lock the shared UI color system so navy is primary, slate is structural, green is success/live, amber is caution, and red is reserved for destructive/error emphasis	Engineering	DONE	Continue the UI lock by converting any remaining page-local legacy color classes onto the shared token system as the next visual recommendations are implemented	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/BrandHeader.test.tsx src/components/PrimaryNavigation.test.tsx src/components/PageArchetypes.test.tsx src/components/FloatingRouterHelper.test.tsx --reporter=dot -> 15 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 97 passed; shared shell grep confirmed no remaining #EE0000, #1a2b56, #243869, or old blue helper classes in frontend/src/components / frontend/src/index.css
T-109	P1	Define and apply four shared page archetype shells (Workspace, Calculator, Catalog, Assistant) so the main tabs stop mixing layout patterns	Engineering	DONE	Extend the same shared shells to the remaining assistant-class tabs (RouterKnowledgebase, RoutersAssistant, MastersAI, PotsAssistant) and then decide whether any shell-specific cleanup is still needed per page	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/PageArchetypes.test.tsx src/components/BrandHeader.test.tsx src/pages/PotsWorkspace.test.tsx --reporter=dot -> 15 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 97 passed; local desktop/mobile browser spot-check confirmed POTS, Telco, Rapid Router, and Knowledgebase render the expected archetype shells
T-108	P1	Consolidate floating support and helper controls into one shared help launcher with internal tabs	Engineering	DONE	Continue the UI lock by reviewing any remaining duplicated top-level utility affordances and deciding whether command-palette/status visibility should stay as-is or be simplified further	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/FloatingRouterHelper.test.tsx src/components/PrimaryNavigation.test.tsx src/components/BrandHeader.test.tsx --reporter=dot -> 12 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 26 files / 94 passed; local desktop/mobile browser spot-check confirmed one floating Help launcher with Assist and Support tabs
T-107	P1	Remove emoji-style workspace cues and standardize the shell on restrained enterprise navigation icons	Engineering	DONE	Continue the UI lock by consolidating the duplicate floating global launchers so the cleaner shell is not undercut by competing bottom-of-screen controls	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/PrimaryNavigation.test.tsx src/components/BrandHeader.test.tsx --reporter=dot -> 8 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 25 files / 90 passed; local desktop/mobile browser spot-check confirmed the workspace shell no longer exposes 🧠 📚 🧮 📉 📡 ⚡ text
T-106	P1	Replace the old toolbox interaction with a true primary navigation system: visible desktop workspace rail, mobile workspace sheet, and integrated workspace search	Engineering	DONE	Continue the UI lock by simplifying the remaining persistent global controls, starting with the duplicate bottom launchers (Get support, Open router helper)	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/PrimaryNavigation.test.tsx src/components/BrandHeader.test.tsx --reporter=dot -> 7 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 25 files / 89 passed; local desktop/mobile browser spot-check confirmed the old Open toolbox / Toolbox is collapsed copy is gone
T-105	P1	Collapse the global shell into one compact utility header and remove always-visible toolbox chrome	Engineering	DONE	Treat the compact header as the locked baseline for the new primary navigation shell; no separate follow-up remains beyond the broader UI lock work	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/BrandHeader.test.tsx --reporter=dot -> 4 passed; cd frontend && npm run build -> success
T-104	P1	Redeploy the Hugging Face hosted frontend so hosted POTS QA runs against the latest simplified workspace/intake implementation instead of the stale stacked build	Engineering	IN_PROGRESS	Trigger/reconfirm the Space rebuild, then rerun the hosted POTS desktop/mobile sign-off pass against the refreshed deployment	Hosted/Auth0 check on 2026-03-06: cd frontend && npx playwright test e2e/auth.full-flow.spec.ts --config=playwright.config.ts --reporter=line -> 1 passed; hosted POTS desktop/mobile inspection -> 0/2 sign-off passes because both viewports still rendered the older stacked POTS workspace with POTS Project Workspace, POTS Estimates + Intake, and POTS Savings Estimator on one page
T-103	P1	Sweep remaining frontend destructive actions so saved drafts, chat resets, and scoped removals all require confirmation before data loss	Engineering	DONE	If final sign-off needs it, spot-check representative destructive flows in the hosted/authenticated runtime; local/frontend verification is complete	cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/utils/chatCommands.test.ts src/utils/confirmAction.test.ts src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsEstimateIntake.test.tsx src/pages/PotsIntake.test.tsx src/pages/PotsWorkspace.test.tsx --reporter=dot -> 27 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 24 files / 86 passed
T-102	P1	Simplify PotsWorkspace into a progressive, one-step-at-a-time surface with collapsed secondary tools	Engineering	DONE	If final deploy sign-off needs it, repeat the same flow against the hosted/authenticated runtime; no more default-open changes are pending from the local browser pass	cd frontend && npx vitest run src/pages/PotsWorkspace.test.tsx --reporter=dot -> 7 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 23 files / 79 passed; local browser QA at 1440x1024 + 390x844 confirmed the support panels now behave as a true accordion
T-101	P1	Simplify the active POTS estimate/intake UX with progressive disclosure and fewer always-open support panels	Engineering	DONE	Treat hosted/authenticated browser QA as optional final sign-off only; local browser QA did not justify opening any additional intake disclosures by default	cd frontend && npx vitest run src/pages/PotsIntake.test.tsx src/pages/PotsEstimateIntake.test.tsx --reporter=dot -> 6 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 23 files / 79 passed; local browser QA at 1440x1024 + 390x844 confirmed See all sites, optional notes, and helper disclosures can stay closed by default
T-100	P1	Clarify POTS estimator start-path choices and seed intake according to the chosen entry mode	Engineering	DONE	Run hosted/manual QA on all three chooser paths (quick estimate, totals now, site details next, site-by-site now) and confirm the seeded intake drafts feel obvious under real auth/runtime conditions	cd frontend && npx vitest run src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsEstimateIntake.test.tsx --reporter=dot -> 5 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 22 files / 72 passed
T-099	P1	Add clear project deletion flow to POTS workspace with confirmation pop-up	Engineering	DONE	Use the selector delete action during hosted/manual QA and note any copy/layout polish or additional destructive-action confirmation gaps elsewhere in the app	cd frontend && npx vitest run src/pages/PotsWorkspace.test.tsx --reporter=dot -> 5 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 20 files / 67 passed; python3 -m pytest -q backend/app/test_pots_workspace_api.py -> 46 passed; python3 -m pytest -q backend/app/test_pots_workspace_api.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 51 passed
T-098	P1	Expose phase-9-24 POTS workspace workflow controls in the SPA for manual/hosted verification	Engineering	DONE	Use the new workflow panel for credentialed hosted/browser QA, export review, and responsive checks; continue tracking any remaining phase-25+ surface gaps separately	cd frontend && npx vitest run src/pages/PotsWorkspace.test.tsx --reporter=dot -> 3 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 20 files / 65 passed; python3 -m pytest -q backend/app/test_pots_workspace_api.py -> 45 passed
T-097	P1	Deep-dive Phase 9-40 workflow logic for gotchas and add detailed edge-case regression coverage	Engineering	DONE	Continue hosted manual UX verification for phase-9+ workflow controls and export review with pilot users	python3 -m pytest -q backend/app/test_pots_workspace_api.py -k \"remove_last_location_resets_project_counts or excel_export_has_required_tabs\" -> 2 passed; python3 -m pytest -q backend/app/test_pots_workspace_api.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 50 passed; cd backend && python3 -m pytest -q -> 435 passed; npm --prefix frontend run build -> success; npm --prefix frontend run test -> 19 files / 62 passed
T-096	P1	Execute Phases 9-40 of POTS workspace roadmap in strict order (guided discovery through launch optimization)	Engineering	DONE	Core backend roadmap and initial SPA workflow surface are complete; next step is hosted/manual UX pass plus credentialed browser journeys using the new PotsWorkspace controls	Per-phase gate: for p in $(seq 9 40); do python3 -m pytest -q backend/app/test_pots_workspace_api.py -k \"phase${p}\"; done -> each selector 1 passed; consolidated: python3 -m pytest -q backend/app/test_pots_workspace_api.py -> 43 passed; python3 -m pytest -q backend/app/test_pots_workspace_api.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 48 passed; npm --prefix frontend run build -> success; npm --prefix frontend run test -> 19 files / 62 passed
T-095	P1	Execute Phase 8 of 40-phase POTS roadmap: audit log v1	Engineering	DONE	Start Phase 9 guided-discovery question tree implementation	python3 -m pytest -q backend/app/test_pots_workspace_api.py -> 11 passed; python3 -m pytest -q backend/app/test_pots_workspace_api.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 16 passed
T-094	P1	Execute Phase 7 of 40-phase POTS roadmap: delegation skeleton (internal section assignment)	Engineering	DONE	Start Phase 8 audit log v1 for immutable activity timeline	python3 -m pytest -q backend/app/test_pots_workspace_api.py -> 10 passed; python3 -m pytest -q backend/app/test_pots_workspace_api.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 15 passed
T-093	P1	Execute Phase 6 of 40-phase POTS roadmap: intake progress model and completion scoring	Engineering	DONE	Start Phase 7 delegation skeleton (internal assignment ownership)	python3 -m pytest -q backend/app/test_pots_workspace_api.py -> 9 passed; python3 -m pytest -q backend/app/test_pots_workspace_api.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 14 passed
T-092	P1	Execute Phase 5 of 40-phase POTS roadmap: workspace-home UX (mode-first start + next action guidance)	Engineering	IN_REVIEW	Complete manual desktop/tablet/mobile QA checklist in docs/dev/pots_workspace_phase5_home_ux.md and close remaining layout nits	npm --prefix frontend run build -> success; npm --prefix frontend run test -> 19 files / 62 passed
T-091	P1	Execute Phase 4 of 40-phase POTS roadmap: tenant/user isolation hardening	Engineering	DONE	Start Phase 5 workspace-home UX refinement and next-action card design	python3 -m pytest -q backend/app/test_pots_workspace_api.py -> 8 passed; python3 -m pytest -q backend/app/test_pots_workspace_api.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 13 passed
T-090	P1	Execute Phase 3 of 40-phase POTS roadmap: lifecycle state machine with guarded transitions	Engineering	DONE	Start Phase 4 tenant/user isolation hardening and fallback handling checks	python3 -m pytest -q backend/app/test_pots_workspace_api.py -> 7 passed; python3 -m pytest -q backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 5 passed
T-089	P1	Execute Phase 2 of 40-phase POTS roadmap: role/collaboration model (internal-first)	Engineering	DONE	Use published role matrix to drive delegation/audit implementation phases; external customer contribution remains deferred	`rg -n "Role Matrix
T-088	P1	Execute Phase 1 of the new 40-phase POTS workspace roadmap (scoped project API + triage + workspace shell)	Engineering	DONE	Begin Phase 2 role/collaboration model and external-contribution boundary decisioning (kept deferred from Phase 1 per user direction)	python3 -m pytest -q backend/app/test_pots_workspace_api.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 9 passed; npm --prefix frontend run build -> success; npm --prefix frontend run test -> 19 files / 62 passed
T-087	P1	Publish detailed 40-phase project map for enterprise POTS workspace execution	Engineering	DONE	Use this map as execution baseline; track phase-by-phase completion status in this task table and session handoff	docs/dev/pots_workspace_40_phase_project_map.md created with phase definitions, verification gates, and exit criteria
T-086	P1	Execute saved cross-workstream gameplan for remaining fixes/enhancements (phased next-thread plan)	Engineering	IN_PROGRESS	Keep parser backlog item deferred; auth blocker is cleared; next focus is broader hosted/manual sign-off items plus optional 150 stability push from 94.7% to >=95%	Phase 0 auth refresh: cd frontend && npx vitest run src/auth/config.test.ts src/auth/errorUtils.test.ts src/components/HealthStatusModal.test.tsx -> 16 passed; python3 -m pytest -q backend/app/test_auth.py backend/app/test_startup_rate_limit.py -> 31 passed; cd frontend && npx playwright test e2e/auth.spec.ts --reporter=line -> 6 passed; cd frontend && npx playwright test e2e/auth.full-flow.spec.ts --reporter=line -> 1 passed; Phase 1 gate run: frontend build success + frontend tests 59 passed + Rapid Router/API pytest 49 passed; Phase 2 gate re-run: frontend build success + frontend tests 59 passed + consolidation pytest 68 passed; Phase 3 gate run: 150 142/150 (94.7%) failed [24,36,88,98,99,104,112,129] (docs/evals/20260305T013817_phase3_gate150_final/unified_kb_eval150_shards10_summary.json), 75 74/75 (98.7%) failed [3] (docs/evals/20260305T015614_phase3_gate75_final/unified_kb_eval150_shards10_summary.json), 50 50/50 (100.0%) failed [] (docs/evals/20260305T020530_phase3_gate50_final/unified_kb_eval150_shards10_summary.json); additional 150 target attempt 141/150 (94.0%) failed [48,55,78,89,99,107,110,112,118] (docs/evals/20260305T021154_phase3_gate150_rerun2_final/unified_kb_eval150_shards10_summary.json); Phase 4 gate run: python3 -m pytest -q backend/app/test_unified_kb_core.py backend/app/test_knowledgebase_api.py backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 151 passed; Phase 5 targeted runs: cd backend && python3 -m pytest -q app/test_unified_kb_core.py app/test_pots_conversation_regression.py app/test_unified_kb_eval150_script.py -> 102 passed
T-085	P1	Validate new Smart Profile + Customer Memory + resume/repeat flows across KB, POTS, and Rapid Router in hosted runtime	Engineering	IN_REVIEW	Seed frontend/.env.e2e or frontend/.env.e2e.local with two Auth0 test users, then run cd frontend && npx playwright test e2e/rapid-router.memory-isolation.spec.ts plus the manual same-browser two-user swap: (1) save/apply Rapid Router profile as user A, (2) log out/in as user B and confirm no customer-profile carryover appears, (3) switch back to user A and confirm scoped memory is still present, (4) repeat KB/POTS handoff checks	npm --prefix frontend run build -> success; cd frontend && npx vitest run src/utils/customerMemory.test.ts --pool=threads --maxWorkers=1 -> 4 passed; cd frontend && npx playwright test e2e/rapid-router.memory-isolation.spec.ts --list -> 1 test listed; cd frontend && npx vitest run src/components/BrandHeader.test.tsx --pool=threads --maxWorkers=1 -> 4 passed
T-084	P1	Validate header Slack-chip responsiveness and spacing with command/status toggles on narrow widths	Engineering	IN_REVIEW	Manually check header controls at mobile/tablet/desktop and with command-palette/system-status hidden; ensure Slack chip remains accessible without wrapping collisions	npm --prefix frontend run build -> success; cd frontend && npx vitest run src/components/BrandHeader.test.tsx --pool=threads --maxWorkers=1 -> 4 passed
T-083	P1	Validate global floating support launcher UX in hosted runtime (desktop + mobile overlap with router helper)	Engineering	IN_REVIEW	Capture hosted screenshots and confirm Slack/email/phone links open correctly from each tab; tune spacing/z-index if mobile overlaps with bottom-page controls	npm --prefix frontend run build -> success; manual hosted check from all enabled tabs
T-082	P1	Validate hosted UX for new Rapid Router split-shipping flow (single-model orders only) across desktop/tablet/mobile	Engineering	IN_REVIEW	Run manual hosted pass covering: default single-address flow, enabling split shipping on one selected model, cap enforcement (locations <= qty), and mixed-model disabled state; capture screenshots and any copy/layout nits	npm --prefix frontend run build -> success; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 25 passed; python3 -m pytest -q backend/app/test_rapid_router_api_shell.py -> 24 passed
T-080	P0	Remove legacy masters-toolkit-api audience assumptions from auth-required deployments and verify hosted login without a custom API audience	Engineering	DONE	Legacy audience placeholder is now ignored in active auth code and hosted login passed with credentialed Playwright runs; keep deployment env clean by leaving VITE_AUTH0_AUDIENCE / AUTH0_AUDIENCE unset unless a real Auth0 API Identifier is introduced later	cd frontend && npx vitest run src/auth/config.test.ts src/auth/errorUtils.test.ts src/components/HealthStatusModal.test.tsx -> 16 passed; python3 -m pytest -q backend/app/test_auth.py backend/app/test_startup_rate_limit.py -> 31 passed; npm --prefix frontend run build -> success; cd frontend && npx playwright test e2e/auth.spec.ts --reporter=line -> 6 passed; cd frontend && npx playwright test e2e/auth.full-flow.spec.ts --reporter=line -> 1 passed
T-081	P1	Fill missing Crown (ASKNCM1100E) WAN/LAN detail fields in deterministic router fact CSV for cleaner Dragon-vs-Crown compares	Engineering	DONE	Added source-backed Crown interface counts to deterministic CSV and covered fast-path behavior with regression assertions	python3 -m pytest -q backend/app/test_unified_kb_core.py backend/app/test_knowledgebase_api.py backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 151 passed
T-079	P1	Recover mixed-domain shard regressions from OpenAI 2026-02-27 run (150 suite failed IDs)	Engineering	IN_REVIEW	Latest best Phase-3 150 gate remains 142/150 (94.7%) with failed IDs [24,36,88,98,99,104,112,129]; additional rerun showed semantic variance (141/150, failed [48,55,78,89,99,107,110,112,118]); focus next on masters/pots long-form semantic stability to hold >=95% target	Re-run cd backend && CHUNK_SIZE=15 START_ID=1 END_ID=150 SEMANTIC_POLICY=all OUT_DIR=../docs/evals/<stamp> CASES_PATH=../docs/evals/unified_kb_eval150_cases.json ./scripts/run_unified_kb_eval150_chunks.sh; artifacts: docs/evals/20260305T013817_phase3_gate150_final/ and docs/evals/20260305T021154_phase3_gate150_rerun2_final/; maintain >=92%, target >=95%
T-078	P1	Raise router-helper quality for generated 50-question conceptual set (currently 23/50)	Engineering	DONE	Completed targeted conceptual-intent/routing fixes in backend/app/knowledgebase/core.py; raised generated-50 suite from 23/50 to 47/50 (94.0%) with no stage-budget exits in latest run	cd backend && CHUNK_SIZE=5 START_ID=1 END_ID=50 SEMANTIC_POLICY=all OUT_DIR=../docs/evals/shards10_eval50_openai_all_20260227_fix7_full CASES_PATH=../docs/evals/unified_kb_eval50_new_questions_router_helper_cases.json ./scripts/run_unified_kb_eval150_chunks.sh
T-077	P1	Consolidate Routers tab capabilities into Master’s Telecom AI Knowledgebase as single source tab (no duplicate tool surfaces)	Engineering	IN_REVIEW	Hosted parity sign-off remains: capture credentialed runtime proof for KB-first router journey and final tab-retirement readiness notes	npm --prefix frontend run build -> success; npm --prefix frontend run test -> 19 files / 59 passed; python3 -m pytest -q backend/app/test_knowledgebase_api.py backend/app/routers/router_tab_smoke_test.py backend/app/test_tab_final_pass_matrix.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 68 passed
T-076	P1	Merge POTS Savings Estimator + POTS Replacement Intake into one guided tab with estimator-to-intake handoff	Engineering	IN_REVIEW	Hosted guided-flow sign-off remains: run credentialed journey for estimator->intake carryover and confirm user-facing prefill clarity	npm --prefix frontend run build -> success; npm --prefix frontend run test -> 19 files / 59 passed; python3 -m pytest -q backend/app/test_knowledgebase_api.py backend/app/routers/router_tab_smoke_test.py backend/app/test_tab_final_pass_matrix.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 68 passed
T-075	P1	Run credentialed hosted browser E2E for full tab journeys (real page-to-page progression)	Engineering	IN_REVIEW	Auth-only hosted runs are green; next expand into the new POTS workspace workflow panel plus existing KB/Rapid Router flows and capture screenshots/artifacts	cd frontend && npx playwright test e2e/auth.spec.ts --reporter=line -> 6 passed; cd frontend && npx playwright test e2e/auth.full-flow.spec.ts --reporter=line -> 1 passed
T-074	P1	Implement non-Rapid tab UI polish pack from cross-tab advisory review	Engineering	IN_REVIEW	Phase-1 quick wins and automated deep-dive visual QA are complete (21 viewport-tab runs, 0 visual issues); execute remaining phase-2/phase-3 structural interactions	npm --prefix frontend run build -> success; npm --prefix frontend run test -> 18 files / 54 tests passed; visual audit frontend/frontend/tmp/visual_audit/visual_audit_results.json shows failedRuns=0, totalVisualIssues=0
T-073	P1	Simplify helper comparison-table UX to table-first output with clearer CTA	Engineering	DONE	Added table-detection/simplification in global helper and bypassed long-answer preview/details for table responses; aligned CTA wording across helper table renderers and published checkpoint commit	npm --prefix frontend run build -> success; commit 1014b78; git push origin main + git push hf-fourtab main
T-072	P1	Publish router-ingestion checkpoint to required remotes	Engineering	DONE	Committed and pushed current router RAG mapping/report/doc updates to both required remotes	commit 8050c76; git push origin main -> 21c3962..8050c76; git push hf-fourtab main -> 21c3962..8050c76
T-071	P1	Ingest new router knowledgebase corpus batch (EX400, RX400, ER815, IR624, Balance 310X)	Engineering	DONE	Added deterministic intake mappings and executed full intake pipeline on staged batch source; verified manifest/chunk inclusion and recall smoke pass	bash backend/scripts/router_rag_intake_pipeline.sh ../tmp/router_rag_intake_2026-02-27_batch -> included=7, skipped=0; python3 backend/scripts/router_rag_smoke.py --query ... -> 5 queries, 0 failures
T-068	P1	Add basic CAPTCHA gate before Rapid Router order submit and first Knowledgebase/POTS/helper request	Engineering	DONE	Completed backend challenge/verify APIs + scoped token enforcement + frontend one-time gate cards + regression coverage	python3 -m pytest -q backend/app/test_rapid_router_api_shell.py backend/app/test_knowledgebase_api.py backend/app/test_chat_guidance_api.py backend/app/rapid_router/test_rapid_router_core.py -> 57 passed; npm --prefix frontend run build -> success
T-059	P1	Add Rapid Router CSV ingestion validator + dry-run import path (schema/lint + duplicate/SKU checks + preview)	Engineering	DONE	Completed core CSV validator + duplicate checks + dry-run/apply path + admin API endpoint + regression tests	python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 39 passed
T-067	P1	Execute Rapid Router 10-point readability/simplicity cleanup pass	Engineering	IN_REVIEW	Core 3-phase refactor is implemented in RapidRouter.tsx; run desktop/mobile visual QA and capture any spacing/copy nits before marking done	npm --prefix frontend run build -> success; visual QA checklist pass for step header, staged actions, fix-list-only validation, helper readability, and admin modal flow
T-069	P1	Implement user-requested 12-point Rapid Router + global UI visibility overhaul	Engineering	IN_REVIEW	Deep-dive compliance pass applied final cleanup patches (remove leftover column-focus/copy controls and unify helper compare label to Device details); perform hosted browser QA/deploy smoke	npm --prefix frontend run build -> success; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 45 passed, 9 warnings; python3 -m pytest -q backend/app/test_unified_kb_core.py backend/app/test_knowledgebase_api.py -> 88 passed, 9 warnings
T-070	P1	Run a targeted visual polish sprint for Rapid Router and shared rails/cards	Engineering	IN_REVIEW	Complete hosted desktop/tablet/mobile screenshot QA for the newly shipped polish pass and capture any residual spacing/copy nits before publish	npm --prefix frontend run build -> success; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 45 passed, 9 warnings
T-066	P1	Profile HF Space startup/wake latency with real runtime timings and recommend env tuning	Engineering	IN_REVIEW	Capture startup stage timings from runtime logs/health (bootstrap, csv_sanity, preload, integrity) and decide preload policy (light vs none) for production	HF boot logs include per-stage timing and restart median improves without regressions
T-063	P2	Clean up third-party deprecation warning noise in Rapid Router test runs (reportlab + SWIG/PyMuPDF)	Engineering	DONE	Added narrowly scoped warning filters/containment around vetted external noise while preserving real exception visibility	python3 -m pytest -q backend/app/test_unified_kb_core.py backend/app/test_knowledgebase_api.py backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 151 passed with warning noise contained
T-065	P2	Contain known benign MuPDF startup font-warning noise from seed-doc setup-note extraction	Engineering	DONE	Wrapped setup-note extraction in targeted stderr containment to suppress known benign font spam only	Startup probe python3 - <<'PY' ... RapidRouterCore(...) ... PY now prints clean startup_ok 12 without repeated MuPDF font warning
T-060	P1	Add Rapid Router <-> Knowledgebase catalog sync contract checks	Engineering	DONE	Added contract test asserting seeded Rapid Router catalog remains queryable via KB fast paths and provider wiring	python3 -m pytest -q backend/app/test_unified_kb_core.py backend/app/test_knowledgebase_api.py backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 151 passed
T-061	P1	Add per-stage latency instrumentation and SLO guardrails for KB helper paths	Engineering	DONE	Added per-stage timing fields + stage SLO evaluation in eval script and shard aggregate summary output	cd backend && python3 -m pytest -q app/test_unified_kb_eval150_script.py -> 6 passed
T-062	P1	Strengthen store schema-version migration tests and strict validation for Rapid Router store JSON	Engineering	DONE	Hardened migration/load paths for malformed versions/products/prices and added regression coverage	python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py (included in Phase 4 gate 151 passed)
T-058	P1	Integrate Rapid Router store data into Unified Knowledgebase router-docs answers	Engineering	DONE	Completed provider injection + deterministic Rapid Router fast paths + fallback coverage tests	cd backend && python3 -m pytest -q app/test_unified_kb_core.py app/test_knowledgebase_api.py app/rapid_router/test_rapid_router_core.py -> 92 passed; manual API check of /api/knowledgebase/message with mode=router_docs returned deterministic_rapid_router_catalog_list_fast with rapid_router_store.json sources
T-057	P1	Validate first-login/re-login with real Auth0 credentials in auth-required runtime	Engineering	DONE	Credentialed hosted login/logout verification passed after audience-optional + legacy-placeholder-ignore auth fixes	cd frontend && npx playwright test e2e/auth.full-flow.spec.ts --reporter=line -> 1 passed; cd frontend && npx playwright test e2e/auth.spec.ts --reporter=line -> 6 passed
T-056	P1	Run a focused UX cleanup pass for Rapid Router/toolbox (progressive disclosure + clearer hierarchy)	Engineering	DONE	Completed full 10-item UX pass in one batch (summary rail, completion chips, jump links, table view, review modal, mobile sticky CTA, and helper readability controls)	cd frontend && npm run build; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py; python3 -m pytest -q backend/app/test_rapid_router_api_shell.py
T-055	P0	Implement MSRP + Masters contact dropdown + configuration-options pricing in Rapid Router	Engineering	DONE	Completed and pushed in commit 176ff8f	cd backend && python3 -m pytest app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py app/test_tab_final_pass_matrix.py -q -> 31 passed; cd frontend && npm run build passed
T-054	P2	Track Rapid Router file-size growth during helper rollout	Engineering	IN_REVIEW	Monitor RapidRouter.tsx line growth and decide if helper should be split into dedicated component/module	frontend/src/pages/RapidRouter.tsx line count trend captured across releases
T-053	P1	Add Rapid Router in-page helper chatbot for router selection + rep Q&A	Engineering	IN_REVIEW	Run targeted rep-prompt QA, add feature flag, and add focused frontend tests for helper interactions	Helper responses remain within timeout budget and do not regress Rapid Router submit flow
T-043	P1	Recover local access to backend/app/test_unified_kb_core.py in Dropbox workspace and commit pending local delta	Engineering	DONE	Confirmed file readability and successful targeted pytest execution from Dropbox workspace	wc -l backend/app/test_unified_kb_core.py read succeeds; cd backend && python3 -m pytest -q app/test_unified_kb_core.py -> 93 passed
T-042	P1	Reduce long-tail latency on top 10 slow cases while preserving 150/150 pass rate	Engineering	TODO	Profile and trim delegate/web-fallback on 66,111,86,91,88,85,92,82,93,99 with per-phase budgets and cache hits	Re-run CHUNK_SIZE=10 START_ID=1 END_ID=150 and target p95 < 7000ms, pass=150
T-037	P1	Post-commit stabilization for residual 75-case failure (ID 75)	Engineering	TODO	Reproduce case 75 with profiler traces and patch mixed Verizon gateway + POTS end-to-end response path	docs/evals/shards5_eval75/unified_kb_eval150_shards10_summary.json shows no failed IDs
T-064	P2	Stabilize Rapid Router 25-case suite residual semantic miss (ID 3)	Engineering	TODO	Inspect shards5_rapidrouter25 case 3 response wording and tighten quote-clarification template for W1850 ambiguity without relaxing guardrails	Re-run CHUNK_SIZE=5 START_ID=1 END_ID=25 CASES_PATH=../docs/evals/unified_kb_eval25_rapid_router_cases.json OUT_DIR=../docs/evals/shards5_rapidrouter25 and target 25/25
T-029	P1	Eliminate 75-case p95 regression versus legacy baseline (318.1ms) while preserving pass rate	Engineering	TODO	Profile slow 75 shards (58-64 cluster) and reduce tail in POTS compare/assumption paths	p95 <= 318.1ms target vs docs/evals/shards5_eval75/unified_kb_eval75_shards5_summary.json
T-030	P1	Finalize commit policy for root docs/faq/FAQ_ongoing_candidates.csv churn	Engineering	DONE	Adopted pytest-time isolation policy via backend conftest.py so local regressions default to temp FAQ candidate path unless explicitly overridden	Repeated test runs preserve root FAQ hash (sha256 unchanged before/after targeted pytest)
T-031	P2	Add focused tests for _parallel_index_search budget behavior under slow index calls	Engineering	DONE	Added deterministic slow-stub tests for bounded in-flight submission and shared executor reuse	cd backend && python3 -m pytest -q app/test_unified_kb_core.py -> 93 passed
T-034	P2	Add dedicated latency guard tests for long-form POTS rewrite path in conversation regression suite	Engineering	DONE	Added long-form single-turn and cumulative-turn latency guard tests in POTS regression suite	cd backend && python3 -m pytest -q app/test_pots_conversation_regression.py -> 3 passed
T-036	P1	Clear remaining 75-case failure (ID 75) without degrading other MSRP/Verizon intents	Engineering	TODO	Reproduce case 75, adjust mixed Verizon gateway + POTS synthesis response to satisfy semantic scorer while preserving guardrails	docs/evals/shards5_eval75/unified_kb_eval150_shards10_summary.json shows no failed IDs
T-038	P2	Prevent test-run churn in root docs/faq/FAQ_ongoing_candidates.csv during local regressions	Engineering	DONE	Added session-level fixture that routes FAQ candidate writes to temp path by default in test runs	Root FAQ candidate file hash remains stable after repeat pytest runs under default test env

Backlog

ID	Priority	Task	Owner	Notes
D-240	Added four new required Rapid Router approval attestations (180-day commitment, quote approval before IMEI release, active MDN before shipment, and truth/correctness) with matching frontend + backend validation and persistence	2026-03-07	frontend/src/pages/RapidRouter.tsx, frontend/src/pages/RapidRouter.test.tsx, backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py; cd backend && .venv/bin/python -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py -> 53 passed; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 5 passed; cd frontend && npm run build -> success
B-007	P2	Deferred by instruction: add Paste order lines parser (5 CR602, 2 RX60) to auto-fill quantities/models	Engineering	Explicitly excluded from current execution cycle; revisit only on direct user re-approval
B-005	P2	Add optional cleanup hook for shared search executor in long-lived workers	Engineering	Defensive hardening for unusual shutdown environments
B-006	P2	Add script-level tests for shard runner TREND_FILE/FAQ out-dir isolation defaults	Engineering	Guard against regressions in eval tooling

Done (Recent)

ID	Task	Completed On	Evidence
D-226	Rapid Router review validation links now open the target accordion chain and focus the exact invalid field, fixing closed-accordion jumps in customer/order sections	2026-03-07	frontend/src/pages/RapidRouter.tsx, frontend/src/pages/RapidRouter.test.tsx; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 8 passed; cd frontend && npm run build -> success
D-251	Pruned non-canonical timestamped eval history and local workspace clutter while preserving the canonical eval baselines, runner assets, and cleanup policy docs	2026-03-07	backend/scripts/cleanup_repo_artifacts.py, docs/evals/README.md; python3 backend/scripts/cleanup_repo_artifacts.py --no-backup -> removed_dirs=75, removed_files=62; python3 backend/scripts/cleanup_repo_artifacts.py --dry-run --no-backup -> 0 pending removals; canonical probes for latest_eval25_guarded_gpt_check, latest_eval50_guarded_gpt_check, latest_eval6_concept_check, release_gate, shards10, and shards5_eval75 all returned OK
D-233	Split the remaining dirty worktree into auditable cleanup batches, refreshed the stale Rapid Router final-pass matrix fixture, reran the full backend suite clean, normalized visible-copy casing on remaining active frontend surfaces, and archived timestamped eval reruns outside the repo	2026-03-07	backend/app/knowledgebase/core.py, backend/app/router_rag/core.py, backend/app/test_router_rag_module.py, backend/app/test_tab_final_pass_matrix.py, backend/app/test_unified_kb_core.py, frontend/src/components/PromptCoach.tsx, frontend/src/pages/MastersAI.tsx, frontend/src/pages/PotsIntake.tsx, frontend/src/pages/PotsSavingsEstimator.tsx, frontend/src/pages/RouterKnowledgebase.tsx, frontend/src/pages/RoutersAssistant.tsx, frontend/src/pages/TelcoCalculator.tsx, frontend/src/pages/TelcoCalculator.test.tsx; cd backend && .venv/bin/python -m pytest -q app/test_tab_final_pass_matrix.py -k rapid_router_final_pass_30_case_matrix -> 1 passed; cd backend && bash scripts/test_backend.sh --full -> 523 passed; cd frontend && npx vitest run src/pages/TelcoCalculator.test.tsx --reporter=dot -> 2 passed; cd frontend && npm run build -> success; timestamped eval reruns archived at /Users/petedunn/Desktop/codex_eval_archives/cleanup_eval_artifacts_20260307.tar.gz
D-241	Fixed Rapid Router order-options completion so advanced notes are optional when at least one advanced checkbox is selected, matching backend validation and removing the false review blocker	2026-03-07	frontend/src/pages/RapidRouter.tsx, frontend/src/pages/RapidRouter.test.tsx; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 7 passed; cd frontend && npm run build -> success; cd backend && .venv/bin/python -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py -> 54 passed
D-242	Normalized visible capitalization across active frontend surfaces so form labels/actions use sentence case by default and title case is reserved for structural headings/proper nouns	2026-03-07	frontend/src/components/PromptCoach.tsx, frontend/src/pages/RapidRouter.tsx, frontend/src/pages/TelcoCalculator.tsx, frontend/src/pages/PotsSavingsEstimator.tsx, frontend/src/pages/PotsIntake.tsx, frontend/src/pages/UnifiedKnowledgebase.tsx, frontend/src/pages/RouterKnowledgebase.tsx, frontend/src/pages/MastersAI.tsx, frontend/src/pages/PotsAssistant.tsx, frontend/src/pages/RoutersAssistant.tsx; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx src/pages/TelcoCalculator.test.tsx src/components/FloatingRouterHelper.test.tsx --reporter=dot -> 13 passed; cd frontend && npm run build -> success
D-225	Rapid Router review validation links now open the target accordion and focus the exact invalid field, fixing the closed-accordion navigation bug in customer/order sections	2026-03-07	frontend/src/pages/RapidRouter.tsx, frontend/src/pages/RapidRouter.test.tsx; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 7 passed; cd frontend && npm run build -> success
D-243	Rapid Router advanced configuration notes now become optional when any advanced task checkbox is selected; notes remain required only for advanced requests with no selected checkbox option	2026-03-07	frontend/src/pages/RapidRouter.tsx, frontend/src/pages/RapidRouter.test.tsx, backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py; cd backend && .venv/bin/python -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py -> 53 passed; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 6 passed; cd frontend && npm run build -> success
D-250	Updated Rapid Router so the flow starts on Browse, defaults payment to BoBo, requires a 7-digit BoBo Bill-to phone, and requires customer-information authorization/communication consent plus an authorization-provider name before submit; synced backend order normalization/PDF/email output to persist those fields	2026-03-07	frontend/src/pages/RapidRouter.tsx, frontend/src/pages/RapidRouter.test.tsx, backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py; cd backend && .venv/bin/python -m pytest -x -vv app/rapid_router/test_rapid_router_core.py -> 28 passed; cd backend && .venv/bin/python -m pytest -q app/test_rapid_router_api_shell.py -> 24 passed; frontend tsc passed but frontend Vitest/build stalled after startup under the current unified-exec saturation
D-223	Added shared preferred-public-source guidance to every active server-side web-assisted assistant path so web fallback now explicitly prefers opendevelopment.verizonwireless.com, masterstelecom.com, and 5gstore.com when relevant	2026-03-07	backend/app/assistant_fallback.py, backend/app/knowledgebase/core.py, backend/app/router_rag/core.py, backend/app/masters_ai/core.py, backend/app/pots_ai/core.py; python3 -m py_compile backend/app/assistant_fallback.py backend/app/router_rag/core.py backend/app/masters_ai/core.py backend/app/pots_ai/core.py backend/app/knowledgebase/core.py backend/app/test_router_rag_module.py backend/app/test_unified_kb_core.py backend/app/test_masters_conversation_regression.py backend/app/test_pots_conversation_regression.py -> success; direct smoke -> SMOKE_OK
D-238	Optimized the three dominant broad-suite latency buckets enough to restore 75/75 and 150/150 accuracy with zero stage-budget exits, while leaving a smaller deterministic tail-latency cleanup open for a follow-up pass	2026-03-07	backend/app/assistant_fallback.py, backend/app/knowledgebase/core.py, backend/app/test_unified_kb_core.py; cd backend && .venv/bin/python -m pytest -q app/test_unified_kb_core.py -k 'masters_contact_center_doc_lookup_prefers_filename_match_without_search or masters_pots_materials_overview_prefers_doc_fast_without_search or pots_use_case_compare_prefers_cached_provider_fast or pots_provider_emphasis_summary_routes_fast or pots_generic_objection_prompt_skips_deep_search or pots_discovery_first_routes_to_concept_fast or router_inventory_audit_skips_concept_preflight or router_gateway_device_type_skips_concept_preflight' -> 8 passed; bash backend/scripts/test_backend.sh --full -> 510 passed; docs/evals/20260307_020040_eval75_guarded_gpt_rerun/unified_kb_eval150_shards10_summary.json -> 75 / 75 passed, stage_budget_exit_rate_pct=0.0; docs/evals/20260307_020040_eval150_guarded_gpt_rerun/unified_kb_eval150_shards10_summary.json -> 150 / 150 passed, stage_budget_exit_rate_pct=0.0
D-239	Added a TTL-backed keyed title cache for Masters mention lookups, proved the cache works under refresh suppression, and confirmed that the remaining 31/32/35/37 latency tail still lives in the delegate path rather than in title rescans	2026-03-07	backend/app/knowledgebase/core.py, backend/app/test_unified_kb_core.py; cd backend && .venv/bin/python -m pytest -q app/test_unified_kb_core.py -k 'masters_securefax_doc_lookup_prefers_file_discovery or masters_securefax_doc_lookup_uses_cached_title_rows_within_refresh_ttl or masters_contact_center_doc_lookup_prefers_filename_match_without_search or masters_pots_materials_overview_prefers_doc_fast_without_search' -> 4 passed; bash backend/scripts/test_backend.sh --full -> 511 passed; docs/evals/20260307_023133_eval150_masters_lookup_slice/unified_kb_eval150_31_37.json -> 7 / 7 passed, avg_latency_ms=2499.04, p95_ms=4383.97
D-237	Fixed the 150 case-133 overblock by narrowing the code-adjudication regex to require code/inspection/AHJ context around approved/approval, and profiled the remaining broad-suite latency clusters before the next 75/150 rerun	2026-03-07	backend/app/knowledgebase/core.py, backend/app/test_unified_kb_core.py; cd backend && .venv/bin/python -m pytest -q app/test_unified_kb_core.py -k 'allows_approved_masters_references_outline_request or blocks_code_adjudication_globally or blocks_exact_current_lead_times_globally or blocks_exact_current_band_support_globally or blocks_exact_current_certification_status_globally or blocks_exact_current_lifecycle_date_globally or blocks_exact_current_availability_globally' -> 7 passed; direct case-133 spot-check returned domain='masters', retrieval_mode='masters_outline_fast', timing_ms.total=4.37; bash backend/scripts/test_backend.sh --full -> 502 passed; Router RAG smoke -> 10 queries / 0 failures
D-236	Reran guarded-GPT 25, 50, 75, and 150 against the current baselines, confirming 25 and 50 are stable while exposing new broad-suite latency tails and one 150 overblock (ID 133) that now define the next cleanup target	2026-03-07	docs/evals/20260307_010031_eval25_guarded_gpt_rerun/unified_kb_eval150_shards10_summary.json -> 25 / 25 passed, p95=499.50ms; docs/evals/20260307_010031_eval50_guarded_gpt_rerun/unified_kb_eval150_shards10_summary.json -> 50 / 50 passed, p95=381.53ms; docs/evals/20260307_010031_eval75_guarded_gpt_rerun/unified_kb_eval150_shards10_summary.json -> 75 / 75 passed, p95=3645.73ms, ab_gate.p95_non_regression=False; docs/evals/20260307_010031_eval150_guarded_gpt_rerun/unified_kb_eval150_shards10_summary.json -> 149 / 150 passed, failed_ids=[133], stage_budget_exit_rate_pct=1.33
D-235	Expanded the guarded-GPT concept pack to 50 reusable questions in 5-question shards, added global early refusals for exact/current risky asks, and validated the broader pack at 50 / 50 passed without latency regression	2026-03-07	backend/scripts/run_unified_kb_eval50_guarded_gpt_chunks.sh, docs/evals/unified_kb_eval50_guarded_gpt_cases.json, backend/app/knowledgebase/core.py, backend/app/assistant_fallback.py, backend/app/test_assistant_fallback.py, backend/app/test_unified_kb_core.py, backend/app/test_masters_conversation_regression.py; cd backend && .venv/bin/python -m pytest -q app/test_assistant_fallback.py app/test_unified_kb_core.py -k 'contact_center or exact_current or code_adjudication or high_risk_code_compliance or lead_times_globally or band_support_globally or certification_status_globally or lifecycle_date_globally or availability_globally' -> 11 passed; set -a && source .env.codex && set +a && cd backend && OUT_DIR=../docs/evals/latest_eval50_guarded_gpt_check ./scripts/run_unified_kb_eval50_guarded_gpt_chunks.sh -> 50 / 50 passed; bash backend/scripts/test_backend.sh --full -> 501 passed
D-234	Implemented Phase 1 and Phase 2 together by hardening blocked-case tests, fixing false-positive regulatory matching, narrowing strict-citation gating for safe concept explainers, and expanding deterministic concept preflight so the reusable 25-case guarded-GPT pack now runs 25 / 25 passed with the POTS concept shard in low-millisecond latency	2026-03-06	backend/app/assistant_fallback.py, backend/app/knowledgebase/core.py, backend/app/pots_ai/core.py, backend/app/test_assistant_fallback.py, backend/app/test_pots_conversation_regression.py, backend/app/test_unified_kb_core.py; cd backend && .venv/bin/python -m pytest -q app/test_assistant_fallback.py app/test_pots_conversation_regression.py app/test_unified_kb_core.py -k 'ul_substring or real_ul_compliance or replacement_plain_english or multisite_stays_internal_fast or dual_pathway or copper_sunset' -> 7 passed; cd backend && .venv/bin/python -m pytest -q app/test_assistant_fallback.py app/test_router_rag_module.py app/test_masters_conversation_regression.py app/test_pots_conversation_regression.py app/test_unified_kb_core.py -> 205 passed; set -a && source .env.codex && set +a && backend/scripts/run_unified_kb_eval25_guarded_gpt_chunks.sh -> 25 / 25 passed; bash backend/scripts/test_backend.sh --full -> 493 passed
D-244	Standardized all active backend/runtime LLM defaults, current env examples, and local repo env pins on gpt-5-mini, fixed the POTS GPT-5 temperature incompatibility, and revalidated the guarded-GPT acceptance pack at 25 / 25 passed	2026-03-06	README.md, backend/.env.test.example, .env.codex, backend/.env.codex, backend/app/main.py, backend/app/chat_nlu.py, backend/app/knowledgebase/core.py, backend/app/router_rag/core.py, backend/app/masters_ai/core.py, backend/app/pots_ai/core.py, backend/app/routers/router_core.py, backend/app/test_pots_conversation_regression.py, backend/scripts/unified_kb_eval150.py, backend/scripts/router_rag_eval50.py, backend/scripts/router_rag_smoke.py, backend/scripts/run_unified_kb_eval150_chunks.sh; cd backend && .venv/bin/python -m pytest -q app/test_pots_conversation_regression.py -k 'concept_fallback_for_generic_pots_question or llm_synthesis_omits_temperature_for_gpt5_models' -> 2 passed; bash backend/scripts/test_backend.sh --full -> 478 passed; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npm run build -> success; cd frontend && npm run test -> 31 files / 111 passed; docs/evals/20260306_230403_eval25_gpt5mini_default/unified_kb_eval150_shards10_summary.json -> 25 / 25 passed
D-245	Added a canonical reusable 25-question guarded-GPT eval pack (5 shards of 5) plus a dedicated shard runner, then stabilized the suite to 24 / 25 passed (96.0%) with only residual case 13 left open	2026-03-06	docs/evals/unified_kb_eval25_guarded_gpt_cases.json, backend/scripts/run_unified_kb_eval25_guarded_gpt_chunks.sh, docs/evals/README.md, docs/evals/latest_eval25_guarded_gpt_check/unified_kb_eval150_shards10_summary.json; targeted reruns for IDs 6,7,8,11,15; final aggregate 24 / 25 passed, failed_ids=[13]
D-246	Added a shared assistant-family concept-fallback module with allow/deny gates, gpt-5-mini model-only fallback, explicit provenance labels, fallback-only +4s budget extension, and GPT+web refinement only when the concept answer still needed current/public information	2026-03-06	backend/app/assistant_fallback.py, backend/app/knowledgebase/core.py, backend/app/router_rag/core.py, backend/app/masters_ai/core.py, backend/app/pots_ai/core.py, backend/app/main.py, frontend/src/utils/chatProvenance.ts, frontend/src/components/chat/ConversationHeader.tsx, frontend/src/pages/UnifiedKnowledgebase.tsx, frontend/src/pages/RouterKnowledgebase.tsx, frontend/src/pages/MastersAI.tsx, frontend/src/pages/PotsAssistant.tsx, frontend/src/pages/RoutersAssistant.tsx; cd backend && .venv/bin/python -m pytest -q app/test_assistant_fallback.py app/test_unified_kb_core.py app/test_router_rag_module.py app/test_masters_conversation_regression.py app/test_pots_conversation_regression.py app/test_chat_guidance_api.py app/test_knowledgebase_api.py -> 202 passed; bash backend/scripts/test_backend.sh --full -> 477 passed; cd frontend && npm run test -> 31 files / 111 passed; docs/evals/latest_eval6_concept_check/unified_kb_eval150_shards10_summary.json -> 6 / 6 passed
D-247	Completed the requested full validation sweep: backend full suite green, frontend full suite green, live Playwright reduced to one hosted POTS provider-coverage miss, and OpenAI shard suites landed at 146/150 (97.3%), 73/75 (97.3%), and 50/50 (100%); also patched local provider-card building to backfill missing providers such as MetTel from indexed evidence mapped back to known files	2026-03-06	backend/app/knowledgebase/core.py, backend/app/test_unified_kb_core.py; bash backend/scripts/test_backend.sh --full -> 459 passed + Router RAG smoke 10/10; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npm run build -> success; cd frontend && npm run test -> 30 files / 106 passed; cd frontend && npx playwright test --config=playwright.config.ts -> 9 passed / 1 failed / 4 skipped; docs/evals/20260306_190557_eval150_rerun/unified_kb_eval150_shards10_summary.json; docs/evals/20260306_192259_eval75_rerun/unified_kb_eval150_shards10_summary.json; docs/evals/20260306_193023_eval50_rerun/unified_kb_eval150_shards10_summary.json; focused regressions 2 passed + 2 passed
D-229	Enforced the current UI-lock scan rules by removing collapsed-state banners, hiding the default header status button, demoting coach/browse actions that competed with the page primary CTA, and consolidating Rapid Router stage progression under the sticky cart	2026-03-06	frontend/src/components/AssistantWorkspace.tsx, frontend/src/components/ConversationalSidePanel.tsx, frontend/src/components/PromptCoach.tsx, frontend/src/components/BrandHeader.tsx, frontend/src/pages/RapidRouter.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/AssistantWorkspace.test.tsx src/components/PromptCoach.test.tsx src/components/BrandHeader.test.tsx src/pages/RapidRouter.test.tsx --reporter=dot -> 11 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 30 files / 105 passed; git diff --check -> success
D-228	Standardized UnifiedKnowledgebase, RouterKnowledgebase, RoutersAssistant, MastersAI, and PotsAssistant on one assistant shell with shared auto-collapsing setup, then added focused setup-panel regression coverage	2026-03-06	frontend/src/components/AssistantWorkspace.tsx, frontend/src/components/AssistantWorkspace.test.tsx, frontend/src/pages/UnifiedKnowledgebase.tsx, frontend/src/pages/RouterKnowledgebase.tsx, frontend/src/pages/RoutersAssistant.tsx, frontend/src/pages/MastersAI.tsx, frontend/src/pages/PotsAssistant.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/components/AssistantWorkspace.test.tsx src/components/PageArchetypes.test.tsx --reporter=dot -> 4 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 30 files / 105 passed
D-227	Rebuilt RapidRouter into a staged commerce flow with one active step at a time, a sticky cart rail, and collapsed Commerce tools, then added focused regression coverage for the new behavior	2026-03-06	frontend/src/pages/RapidRouter.tsx, frontend/src/pages/RapidRouter.test.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/RapidRouter.test.tsx --reporter=dot -> 2 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 29 files / 103 passed
D-226	Collapsed Telco assumptions, what-if mode, diagnostics, quote helpers, scenario JSON/CSV, and assistant coaching into one shared Advanced drawer so the default calculator surface stays on the business flow	2026-03-06	frontend/src/pages/TelcoCalculator.tsx, frontend/src/pages/TelcoCalculator.test.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/TelcoCalculator.test.tsx --reporter=dot -> 2 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 28 files / 101 passed
D-248	Rebuilt TelcoCalculator as a four-step sequence (Locations, Pricing, Results, Export) and added regression coverage for the new step flow	2026-03-06	frontend/src/pages/TelcoCalculator.tsx, frontend/src/pages/TelcoCalculator.test.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/TelcoCalculator.test.tsx --reporter=dot -> 1 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 28 files / 100 passed
D-224	Replaced paragraph-style POTS instructions with a shared three-line StepGuide pattern across the merged estimate/intake flow so each step now states what it does, what is needed now, and what happens next	2026-03-06	frontend/src/components/ui.tsx, frontend/src/pages/PotsEstimateIntake.tsx, frontend/src/pages/PotsSavingsEstimator.tsx, frontend/src/pages/PotsIntake.tsx, frontend/src/pages/PotsEstimateIntake.test.tsx, frontend/src/pages/PotsSavingsEstimator.test.tsx, frontend/src/pages/PotsIntake.test.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsEstimateIntake.test.tsx src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsIntake.test.tsx src/pages/PotsWorkspace.test.tsx --reporter=dot -> 23 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 99 passed
D-249	Flattened the embedded PotsEstimateIntake shell by adding explicit embedded-mode rendering to the merged wrapper, estimator, and intake so the combined flow no longer feels like nested full-page cards	2026-03-06	frontend/src/pages/PotsEstimateIntake.tsx, frontend/src/pages/PotsSavingsEstimator.tsx, frontend/src/pages/PotsIntake.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsEstimateIntake.test.tsx src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsIntake.test.tsx src/pages/PotsWorkspace.test.tsx --reporter=dot -> 23 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 99 passed
D-222	Converted PotsWorkspace routing into a one-question-at-a-time conversation with answer cards, compact Why this matters disclosure, and a final review/edit step while preserving the existing triage payload	2026-03-06	frontend/src/pages/PotsWorkspace.tsx, frontend/src/pages/PotsWorkspace.test.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsWorkspace.test.tsx --reporter=dot -> 10 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 99 passed
D-221	Moved active-project creation and saved-project switching into the Project tools drawer so PotsWorkspace no longer shows setup UI in the main wizard by default	2026-03-06	frontend/src/pages/PotsWorkspace.tsx, frontend/src/pages/PotsWorkspace.test.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsWorkspace.test.tsx --reporter=dot -> 9 passed; cd frontend && npm run build -> success
D-220	Converted PotsWorkspace from a stacked dashboard into a true wizard shell with one active step card plus a focused utilities drawer for routing/intake	2026-03-06	frontend/src/pages/PotsWorkspace.tsx, frontend/src/pages/PotsWorkspace.test.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/pages/PotsWorkspace.test.tsx --reporter=dot -> 8 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 27 files / 97 passed; git diff --check -> success
D-219	Completed the app-wide destructive-action confirmation sweep, added shared confirmation helper + cancel-aware slash resets, and covered the highest-risk cancel paths with focused frontend regression tests	2026-03-06	frontend/src/utils/confirmAction.ts, frontend/src/utils/chatCommands.ts, frontend/src/pages/PotsEstimateIntake.tsx, frontend/src/pages/PotsIntake.tsx, frontend/src/pages/PotsWorkspace.tsx, frontend/src/pages/TelcoCalculator.tsx, frontend/src/pages/RapidRouter.tsx, frontend/src/pages/UnifiedKnowledgebase.tsx, frontend/src/pages/RouterKnowledgebase.tsx, frontend/src/pages/MastersAI.tsx, frontend/src/pages/PotsAssistant.tsx, frontend/src/pages/RoutersAssistant.tsx, frontend/src/components/FloatingRouterHelper.tsx; cd frontend && npx tsc -p tsconfig.json --noEmit -> success; cd frontend && npx vitest run src/utils/chatCommands.test.ts src/utils/confirmAction.test.ts src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsEstimateIntake.test.tsx src/pages/PotsIntake.test.tsx src/pages/PotsWorkspace.test.tsx --reporter=dot -> 27 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 24 files / 86 passed
D-218	Finished the remaining POTS intake density pass, added focused intake/workspace regression coverage, and used desktop/mobile browser QA to justify a true single-open workspace accordion plus closed-by-default intake scope disclosures	2026-03-06	frontend/src/pages/PotsIntake.tsx, frontend/src/pages/PotsIntake.test.tsx, frontend/src/pages/PotsWorkspace.tsx, frontend/src/pages/PotsWorkspace.test.tsx; cd frontend && npx vitest run src/pages/PotsIntake.test.tsx src/pages/PotsEstimateIntake.test.tsx src/pages/PotsWorkspace.test.tsx --reporter=dot -> 13 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 23 files / 79 passed; local Playwright/browser QA at 1440x1024 and 390x844 confirmed the final disclosure defaults
D-217	Simplified the active POTS estimate/intake experience by hiding support chrome behind disclosures, gating estimate inputs behind customer basics, and collapsing the full estimate math until results are requested	2026-03-06	frontend/src/pages/PotsSavingsEstimator.tsx, frontend/src/pages/PotsEstimateIntake.tsx, frontend/src/pages/PotsIntake.tsx, frontend/src/pages/PotsSavingsEstimator.test.tsx; cd frontend && npx vitest run src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsEstimateIntake.test.tsx --reporter=dot -> 6 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 22 files / 73 passed
D-216	Clarified POTS estimator start paths with an explicit three-mode chooser and made intake seeding follow the selected mode (quick estimate, totals now, site-by-site now)	2026-03-06	frontend/src/pages/PotsSavingsEstimator.tsx, frontend/src/pages/PotsSavingsEstimator.test.tsx, frontend/src/pages/PotsEstimateIntake.tsx, frontend/src/pages/PotsEstimateIntake.test.tsx; cd frontend && npx vitest run src/pages/PotsSavingsEstimator.test.tsx src/pages/PotsEstimateIntake.test.tsx --reporter=dot -> 5 passed; cd frontend && npm run build -> success; cd frontend && npm run test -> 22 files / 72 passed
D-215	Verified hosted auth-required runtime after removing legacy masters-toolkit-api audience dependency: smoke suite and full credentialed login/logout flow both passed	2026-03-06	cd frontend && npx playwright test e2e/auth.spec.ts --reporter=line -> 6 passed; cd frontend && npx playwright test e2e/auth.full-flow.spec.ts --reporter=line -> 1 passed
D-213	Added dedicated Playwright coverage for Rapid Router two-user saved-profile isolation and enabled ignored local frontend/.env.e2e(.local) loading for repeatable credentialed hosted runs	2026-03-05	frontend/e2e/rapid-router.memory-isolation.spec.ts, frontend/playwright.config.ts, frontend/e2e.env.template, frontend/package.json; npm --prefix frontend run build -> success; cd frontend && npx playwright test e2e/rapid-router.memory-isolation.spec.ts --list -> 1 test listed
D-214	Removed active reliance on legacy Auth0 audience https://masters-toolkit-api by treating it as invalid/ignored in frontend and backend auth config, and added user-facing callback guidance for the exact Service not found error	2026-03-06	frontend/src/auth/config.ts, frontend/src/auth/errorUtils.ts, frontend/src/auth/config.test.ts, frontend/src/auth/errorUtils.test.ts, backend/app/auth.py, backend/app/test_auth.py; cd frontend && npx vitest run src/auth/config.test.ts src/auth/errorUtils.test.ts src/components/HealthStatusModal.test.tsx -> 16 passed; python3 -m pytest -q backend/app/test_auth.py backend/app/test_startup_rate_limit.py -> 31 passed; npm --prefix frontend run build -> success
D-212	Scoped shared Smart Profile, resume cards, POTS carryover, and Rapid Router repeat-draft memory per authenticated user so customer data is no longer browser-global across logins	2026-03-05	frontend/src/utils/customerMemory.ts, frontend/src/utils/customerMemory.test.ts, frontend/src/auth/AuthGate.tsx, frontend/src/main.tsx, frontend/src/pages/RapidRouter.tsx; npm --prefix frontend run build -> success; cd frontend && npx vitest run src/utils/customerMemory.test.ts --pool=threads --maxWorkers=1 -> 4 passed
D-211	Fixed battery-router shortlist omission so removable option (CR202-Lite) is preserved for best routers with batteries and added regression coverage	2026-03-05	backend/app/knowledgebase/core.py, backend/app/test_unified_kb_core.py; PYTHONPATH=backend python3 -m pytest -q backend/app/test_unified_kb_core.py -k "battery_best_list_keeps_removable_option" -> 1 passed; runtime harness probe now returns CR202-Lite in battery options table
D-210	Executed additional 150 rerun attempt to reach >=95% target and logged stochastic variance outcome for follow-up (T-079)	2026-03-05	cd backend && CHUNK_SIZE=15 START_ID=1 END_ID=150 SEMANTIC_POLICY=all OUT_DIR=../docs/evals/20260305T021154_phase3_gate150_rerun2_final CASES_PATH=../docs/evals/unified_kb_eval150_cases.json ./scripts/run_unified_kb_eval150_chunks.sh -> 141/150 (94.0%), failed IDs [48,55,78,89,99,107,110,112,118], stage_budget_exits=0
D-209	Completed gameplan Phase 3 evaluation verification gate command set (150/75/50) with quality floor maintained and docs/eval artifacts published	2026-03-05	150: 142/150 (94.7%), failed [24,36,88,98,99,104,112,129] (docs/evals/20260305T013817_phase3_gate150_final/unified_kb_eval150_shards10_summary.json); 75: 74/75 (98.7%), failed [3] (docs/evals/20260305T015614_phase3_gate75_final/unified_kb_eval150_shards10_summary.json); 50: 50/50 (100.0%), failed [] (docs/evals/20260305T020530_phase3_gate50_final/unified_kb_eval150_shards10_summary.json)
D-208	Completed gameplan Phase 2 consolidation verification gate and moved T-076/T-077 to hosted sign-off (IN_REVIEW)	2026-03-05	npm --prefix frontend run build -> success; npm --prefix frontend run test -> 19 files / 59 tests passed; python3 -m pytest -q backend/app/test_knowledgebase_api.py backend/app/routers/router_tab_smoke_test.py backend/app/test_tab_final_pass_matrix.py backend/app/test_pots_response_contract.py backend/app/test_pots_conversation_regression.py -> 68 passed
D-207	Completed gameplan Phase 5 repo/tooling hygiene hardening (T-030, T-038, T-031, T-034, T-043) with verification reruns	2026-03-05	Added backend/app/conftest.py fixture isolation + warning filters; added _parallel_index_search budget tests and POTS long-form latency guard tests; verified cd backend && python3 -m pytest -q app/test_unified_kb_core.py app/test_pots_conversation_regression.py app/test_unified_kb_eval150_script.py -> 102 passed; root FAQ hash unchanged across repeat run
D-206	Completed gameplan Phase 4 data/contract/migration hardening (T-081, T-060, T-061, T-062, T-063, T-065)	2026-03-05	Updated Crown deterministic facts in feb2026routers.csv; added Rapid Router/KB contract tests and store migration hardening/tests; added stage-timing SLO outputs in eval scripts; contained MuPDF/reportlab noise; phase gate python3 -m pytest -q backend/app/test_unified_kb_core.py backend/app/test_knowledgebase_api.py backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 151 passed
D-205	Completed Phase 1 verification gate from the saved gameplan (frontend build/test + Rapid Router backend regression command set)	2026-03-05	npm --prefix frontend run build -> success; npm --prefix frontend run test -> 19 files / 59 tests passed; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 49 passed, 9 warnings
D-204	Completed Phase 0 auth verification gate from the saved gameplan with hosted URL substitution and explicit credential-blocker capture	2026-03-05	cd frontend && npx vitest run src/auth/config.test.ts src/auth/errorUtils.test.ts -> 13 passed; python3 -m pytest -q backend/app/test_auth.py -> 21 passed; cd frontend && E2E_DISABLE_WEBSERVER=true E2E_BASE_URL=https://crazycrazypete-masters-four-tab-openai.hf.space npx playwright test e2e/auth.full-flow.spec.ts -> 1 skipped
D-203	Saved next-thread phased execution gameplan for remaining fixes/enhancements and recorded explicit parser deferral scope	2026-03-04	docs/dev/next_thread_remaining_fixes_enhancements_gameplan.md; T-086 added; B-007 parser deferred
D-202	Re-ran post-edit verification gate for Smart Profile/customer-memory + resume/carryover + KB action-chip rollout before handoff	2026-03-04	git status --short; npm --prefix frontend run build -> success; cd frontend && npx vitest run src/utils/customerMemory.test.ts --pool=threads --maxWorkers=1 -> 3 passed
D-201	Implemented shared Smart Profile/Customer Memory utility + resume/repeat work cards, hardened estimate->intake carryover replay, and Knowledgebase action chips to Router Helper / Rapid Router order draft	2026-03-04	frontend/src/utils/customerMemory.ts, frontend/src/utils/customerMemory.test.ts, frontend/src/pages/PotsSavingsEstimator.tsx, frontend/src/pages/PotsEstimateIntake.tsx, frontend/src/pages/UnifiedKnowledgebase.tsx, frontend/src/pages/RapidRouter.tsx, frontend/src/App.tsx; npm --prefix frontend run build -> success; cd frontend && npx vitest run src/utils/customerMemory.test.ts --pool=threads --maxWorkers=1 -> 3 passed
D-200	Published consolidated checkpoint commit to both required remotes (origin, hf-fourtab)	2026-03-04	commit fcd2934; git push origin main -> e1ec24c..fcd2934; git push hf-fourtab main -> e1ec24c..fcd2934
D-199	Added persistent header one-click Slack support chip in shared BrandHeader (global across tabs)	2026-03-04	frontend/src/components/BrandHeader.tsx, frontend/src/components/BrandHeader.test.tsx; npm --prefix frontend run build -> success; cd frontend && npx vitest run src/components/BrandHeader.test.tsx --pool=threads --maxWorkers=1 -> 4 passed
D-198	Added global floating support launcher with Slack-first CTA plus email/phone one-click fallback and command-palette open action	2026-03-04	frontend/src/components/FloatingSupportLauncher.tsx, frontend/src/App.tsx; npm --prefix frontend run build -> success; cd frontend && npx vitest run src/components/BrandHeader.test.tsx src/components/PromptCoach.test.tsx --pool=threads --maxWorkers=1 -> 5 passed
D-197	Implemented Rapid Router split-shipping locations for single-model orders with frontend + backend validation and order PDF/email location breakdown	2026-03-04	frontend/src/pages/RapidRouter.tsx; backend/app/rapid_router/core.py; backend/app/rapid_router/test_rapid_router_core.py; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 25 passed, 6 warnings; python3 -m pytest -q backend/app/test_rapid_router_api_shell.py -> 24 passed, 9 warnings; npm --prefix frontend run build -> success
D-196	Imported Dragon quick guide + Spark/Kadet docs into canonical Router RAG corpus and added deterministic Dragon/M106/M519/K500A/K300NB router-fact coverage with alias normalization	2026-03-04	Intake run backend/scripts/router_rag_intake_pipeline.sh /tmp/router_rag_intake_2026-03-04_dragon_spark_kadet -> included=6; canonical files under _RAG_Ready_KB_Organized/01_documents/routers/connect_csg/ and /routers/verizon/; updated feb2026routers.csv; alias/routing updates in backend/app/knowledgebase/core.py; regression python3 -m pytest -q backend/app/test_unified_kb_core.py -k dragon_and_katalyst_phrase_aliases -> 1 passed
D-195	Extracted and reported exact failed-question lists (ID + query text) for recovered 150/75/50 suites from shard artifacts	2026-02-28	Parsed results[] where pass=false from docs/evals/shards15_eval150_openai_all_20260227_fix12/, docs/evals/shards10_eval75_openai_all_20260227_fix8/, docs/evals/shards10_eval50_openai_all_20260227_fix7_full/; counts 8, 2, 3 respectively
D-194	Restored eval quality gate above 92% across all requested OpenAI suites and validated with targeted KB regressions	2026-02-27	python3 -m pytest -q backend/app/test_unified_kb_core.py backend/app/test_knowledgebase_api.py -> 96 passed, 9 warnings; docs/evals/shards15_eval150_openai_all_20260227_fix12/unified_kb_eval150_shards10_summary.json -> 142/150 (94.7%); docs/evals/shards10_eval75_openai_all_20260227_fix8/unified_kb_eval150_shards10_summary.json -> 73/75 (97.3%); docs/evals/shards10_eval50_openai_all_20260227_fix7_full/unified_kb_eval150_shards10_summary.json -> 47/50 (94.0%)
D-193	Patched Auth0 audience normalization to prefer non-trailing-slash identifier first, fixing login callback failures caused by https://masters-toolkit-api/ audience selection	2026-02-27	frontend/src/auth/config.ts, frontend/src/auth/config.test.ts, backend/app/auth.py, backend/app/test_auth.py; cd frontend && npx vitest run src/auth/config.test.ts src/auth/errorUtils.test.ts -> 13 passed; python3 -m pytest -q backend/app/test_auth.py -> 21 passed; npm --prefix frontend run build -> success
D-192	Executed requested OpenAI shard evaluation batch (150 + 75 + new 50) in 10 groups each and published aggregate summaries + failed-ID lists	2026-02-27	docs/evals/shards10_eval150_openai_all_20260227/unified_kb_eval150_shards10_summary.json -> 119/150; docs/evals/shards10_eval75_openai_all_20260227/unified_kb_eval150_shards10_summary.json -> 73/75; docs/evals/shards10_eval50_openai_all_20260227/unified_kb_eval150_shards10_summary.json -> 23/50; generated case pack docs/evals/unified_kb_eval50_new_questions_router_helper_cases.json
D-191	Implemented first-pass single-workspace convergence: added unified POTS Estimates + Intake tab flow with estimator handoff + fresh-start session behavior, plus Knowledgebase action/global command to launch floating router helper from any page	2026-02-27	frontend/src/pages/PotsEstimateIntake.tsx, frontend/src/pages/PotsSavingsEstimator.tsx, frontend/src/App.tsx, frontend/src/pages/UnifiedKnowledgebase.tsx, frontend/src/components/FloatingRouterHelper.tsx; npm --prefix frontend run build -> success; npm --prefix frontend run test -> 18 files / 54 tests passed; python3 -m pytest -q backend/app/test_tab_final_pass_matrix.py backend/app/test_knowledgebase_api.py backend/app/routers/router_tab_smoke_test.py -> 63 passed, 9 warnings
D-190	Completed local cross-tab validation sweep and fixed two test roadblocks (router compare fallback scenario stability + fast E2E skip on non-frontend base URLs)	2026-02-27	python3 -m pytest -q backend/app -> 357 passed; python3 -m pytest -q backend/app/test_tab_final_pass_matrix.py -> 4 passed; npm --prefix frontend run test -> 18 files / 54 tests; BASE_URL=http://127.0.0.1:4173/ node frontend/tmp/visual_audit/run_visual_audit.mjs -> 21 runs, 0 issues; backend/app/routers/router_tab_smoke_test.py; frontend/e2e/upload.features.spec.ts
D-189	Removed recommended wording from Knowledgebase Mode options copy (Auto line) while preserving mode behavior text	2026-02-27	frontend/src/pages/UnifiedKnowledgebase.tsx; npm --prefix frontend run build -> success; npm --prefix frontend run test -> 18 files / 54 tests passed
D-188	Consolidated Knowledgebase answer metadata (Why, Next action, Files, Sources) into a single collapsed Response details accordion	2026-02-27	frontend/src/pages/UnifiedKnowledgebase.tsx; npm --prefix frontend run build -> success; npm --prefix frontend run test -> 18 files / 54 tests passed
D-187	Imported IR302 doc batch (quick guide/user manual/spec), rebuilt chunks, and added deterministic IR302 fact row with MSRP $179.00	2026-02-27	backend/scripts/router_rag_import_corpus.py mapping additions; bash backend/scripts/router_rag_intake_pipeline.sh tmp/router_rag_intake_2026-02-27_ir302 -> included=3; reports docs/reports/router_rag_intake_ir302_20260227TIR302.csv/.md; feb2026routers.csv IR302 row; API probe (router_docs) returned deterministic_router_fact_index with MSRP $179.00
D-186	Verified RV50X Feb2022 -F upload is already in canonical corpus via duplicate-hash mapping and added deterministic RV50X host-interface fact coverage (single Ethernet + RS-232 serial) with regression test	2026-02-27	python3 backend/scripts/router_rag_import_corpus.py --source-dir /tmp/rv50x-intake-* --data-dir _RAG_Ready_KB_Organized ... -> skipped=1 (duplicate_hash -> Semtech-RV50X-Data Sheet-Feb2022.pdf); feb2026routers.csv RV50X row added; backend/app/test_unified_kb_core.py new RV50X host-interface test; python3 -m pytest -q backend/app/test_unified_kb_core.py -k \"router_fact_fast_path_from_csv or rv50x_host_interfaces_include_single_ethernet_and_serial\" -> 2 passed; python3 -m pytest -q backend/app/test_knowledgebase_api.py -> 7 passed, 9 warnings
D-185	Generated and executed an ungraded 50-question Knowledgebase batch and captured full raw responses for manual review/scoring	2026-02-27	docs/evals/kb_50_new_questions_results_2026-02-27.json, docs/evals/kb_50_new_questions_results_2026-02-27.md; python3 - <<'PY' ... FastAPI TestClient batch ... PY -> 50/50 HTTP 200 in ~16.0s
D-184	Replaced Rapid Router primary logo asset with user-provided arrow-logo variant (asset-only swap; no layout logic changes)	2026-02-27	frontend/public/rapid-router-primary-logo.png; source /Users/petedunn/Library/Containers/com.apple.Preview/Data/tmp/PreviewTemp-QpJOdK/Untitled Image 3.png; npm --prefix frontend run build -> success
D-183	Completed deep-dive multi-viewport render verification and patched residual overflow hotspots (BrandHeader mobile wrapping, Rapid setup-note URL wrapping, Rapid signature block overflow containment)	2026-02-27	frontend/src/components/BrandHeader.tsx, frontend/src/pages/RapidRouter.tsx, frontend/src/pages/UnifiedKnowledgebase.tsx, frontend/src/pages/RouterKnowledgebase.tsx, frontend/src/pages/RoutersAssistant.tsx; npm --prefix frontend run build -> success; npm --prefix frontend run test -> 54 passed; visual audit frontend/frontend/tmp/visual_audit/visual_audit_results.json -> 21 runs, 0 failures, 0 visual issues
D-182	Executed phase-1 non-Rapid cross-tab UI polish pass (shared chat table renderer + sticky composers + Telco table readability + POTS side-rail/flow quick wins)	2026-02-27	frontend/src/components/chat/markdownTableComponents.tsx, frontend/src/components/chat/ChatComposer.tsx, frontend/src/pages/UnifiedKnowledgebase.tsx, frontend/src/pages/RouterKnowledgebase.tsx, frontend/src/pages/RoutersAssistant.tsx, frontend/src/pages/TelcoCalculator.tsx, frontend/src/pages/PotsSavingsEstimator.tsx, frontend/src/pages/PotsIntake.tsx; npm --prefix frontend run build -> success
D-181	Added centered Rapid Router header logo block using new public asset and responsive framed hero treatment	2026-02-27	frontend/src/pages/RapidRouter.tsx, frontend/public/rapid-router-primary-logo.png; npm --prefix frontend run build -> success
D-180	Completed non-Rapid tab UI/visual advisory sweep and produced per-tab advanced suggestion pack (no-code)	2026-02-27	Reviewed frontend/src/App.tsx + non-Rapid page components; recommendations delivered in chat; no runtime code changes
D-179	Published helper comparison-table UX simplification checkpoint to both required remotes	2026-02-27	commit 1014b78; git push origin main -> 087d265..1014b78; git push hf-fourtab main -> 087d265..1014b78
D-178	Simplified helper comparison output to table-first UI and made CTA consistently explicit (Click here for comparison table)	2026-02-27	frontend/src/components/FloatingRouterHelper.tsx, frontend/src/pages/RapidRouter.tsx; npm --prefix frontend run build -> success
D-177	Published router-ingestion checkpoint commit to both required remotes	2026-02-27	commit 8050c76; git push origin main -> 21c3962..8050c76; git push hf-fourtab main -> 21c3962..8050c76
D-176	Processed and ingested 7 new router PDFs into canonical Router RAG corpus with deterministic mapping, parse/chunk rebuild, and recall verification	2026-02-27	backend/scripts/router_rag_import_corpus.py; docs/reports/router_rag_intake_2026-02-27_batch_import_report_20260227T005515Z.csv (included=7, skipped=0); python3 backend/scripts/router_rag_smoke.py --query ... -> 5/5 pass
D-175	Published Rapid Router UI polish/readability checkpoint to both required remotes	2026-02-27	commit ac92a10; git push origin main -> 9897015..ac92a10; git push hf-fourtab main -> 9897015..ac92a10
D-174	Executed full Rapid Router/floating-helper visual polish batch from advisory list (density toggle, staged submit CTA hierarchy, compact right rail, expandable fix list, helper long-answer details)	2026-02-27	frontend/src/pages/RapidRouter.tsx, frontend/src/components/FloatingRouterHelper.tsx; npm --prefix frontend run build -> success; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 45 passed, 9 warnings
D-173	Published generalized Ericsson/CradlePoint ...50 non-WiFi alias mapping checkpoint to both required remotes	2026-02-26	commit b3420ef; git push origin main -> aa0ddb8..b3420ef; git push hf-fourtab main -> aa0ddb8..b3420ef
D-172	Generalized Ericsson/CradlePoint ...50 model alias rule so non-WiFi variants map to ...00 base models (for example AER2250 -> AER2200) with matching variant notes/Wi-Fi override behavior	2026-02-26	backend/app/routers/router_core.py, backend/app/routers/router_tab_smoke_test.py; python3 -m pytest -q backend/app/routers/router_tab_smoke_test.py -> 52 passed, 9 warnings
D-171	Published Rapid Router rail-width + currency-alignment patch checkpoint to both required remotes	2026-02-26	commit 00ea9d8; git push origin main -> 70f3a5c..00ea9d8; git push hf-fourtab main -> 70f3a5c..00ea9d8
D-170	Tightened Rapid Router right-rail width and hardened per-card currency rendering with split $ + amount columns for stable symbol alignment in both pricing and unit/subtotal blocks	2026-02-26	frontend/src/pages/RapidRouter.tsx; npm --prefix frontend run build -> success
D-169	Unblocked POTS Intake line-inventory spreadsheet flow by enabling Keep number / port needed? toggles while keeping requirement enforcement	2026-02-26	frontend/src/pages/PotsIntake.tsx; npm --prefix frontend run build -> success
D-168	Fixed Routers typo/parse reliability for inventory pastes: 12 RX60 no longer misparses as 12 R x60, and likely transposed typo models now trigger confirmation (RX60 -> XR60) before snapshot execution	2026-02-26	backend/app/routers/router_core.py, backend/app/routers/router_tab_smoke_test.py; python3 -m pytest -q backend/app/routers/router_tab_smoke_test.py -> 50 passed, 9 warnings
D-167	Fixed Routers inventory customer ownership carry-forward for Customer has qty model, qty model... syntax and added regression coverage for Hoover has 200 IBR650, 12 AER2200, 16 MG51	2026-02-26	backend/app/routers/router_core.py, backend/app/routers/router_tab_smoke_test.py; python3 -m pytest -q backend/app/routers/router_tab_smoke_test.py -> 47 passed, 9 warnings
D-166	Rebalanced Rapid Router column layout (narrower right rail, wider left router cards) and aligned dollar signs in both top pricing and Unit/Subtotal blocks	2026-02-26	frontend/src/pages/RapidRouter.tsx; npm --prefix frontend run build -> success
D-165	Committed and pushed dollar-sign alignment fix for Rapid Router pricing rows to both required remotes	2026-02-26	commit ae70744; git push origin main -> 8584959..ae70744; git push hf-fourtab main -> 8584959..ae70744
D-164	Implemented explicit dollar-sign vertical alignment in product-card pricing by using a shared fixed-width value column and left-aligned currency strings	2026-02-26	frontend/src/pages/RapidRouter.tsx; npm --prefix frontend run build -> success
D-163	Committed and pushed follow-up laptop-width pricing-readability hardening to both required remotes	2026-02-26	commit 6312e7d; git push origin main -> fa21c6f..6312e7d; git push hf-fourtab main -> fa21c6f..6312e7d
D-162	Added follow-up pricing readability hardening: xl card density reduced to 3 columns and price rows use fixed value-column width for cleaner label/value separation	2026-02-26	frontend/src/pages/RapidRouter.tsx; npm --prefix frontend run build -> success
D-161	Committed and pushed pricing-overlap readability hotfix to both required remotes	2026-02-26	commit dfd9f34; git push origin main -> 07fc56e..dfd9f34; git push hf-fourtab main -> 07fc56e..dfd9f34
D-160	Fixed Rapid Router product-card pricing readability by replacing overlapping compact pricing grid with row-based flex layout (MSRP, Standard FWA, Backup / Pooled)	2026-02-26	frontend/src/pages/RapidRouter.tsx; npm --prefix frontend run build -> success
D-159	Committed and pushed deep-dive Rapid Router/helper compliance checkpoint to both required remotes	2026-02-26	commit 2f4082e; git push origin main -> a957b5c..2f4082e; git push hf-fourtab main -> a957b5c..2f4082e
D-158	Ran deep-dive visual-compliance code audit for prior Rapid Router/helper requests and patched two remaining inconsistencies: removed legacy Column focus/Copy CSV controls from remaining table-reader path and unified generic compare label to Device details	2026-02-26	frontend/src/pages/RapidRouter.tsx, backend/app/knowledgebase/core.py; npm --prefix frontend run build -> success; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 45 passed; python3 -m pytest -q backend/app/test_unified_kb_core.py backend/app/test_knowledgebase_api.py -> 88 passed
D-157	Added global floating Router helper across all tabs, moved/kept Rapid Router Find+Filter in right rail above Order status, enforced runtime HF visibility flags for admin/command-palette/system-status, and preserved Activation verification as top/default configuration option	2026-02-26	frontend/src/components/FloatingRouterHelper.tsx, frontend/src/App.tsx, frontend/src/components/BrandHeader.tsx, frontend/src/pages/RapidRouter.tsx, backend/app/main.py; npm --prefix frontend run build -> success; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 45 passed; python3 -m pytest -q backend/app/test_unified_kb_core.py backend/app/test_knowledgebase_api.py -> 88 passed
D-156	Hardened Auth0 token finalization flow to prevent premature timeout/config errors during active token setup; added preferred-audience persistence/rotation and timeout-specific guidance	2026-02-26	frontend/src/auth/AuthGate.tsx; npm --prefix frontend run build -> success; cd frontend && npx vitest run src/auth/config.test.ts src/auth/errorUtils.test.ts -> 12 passed; python3 -m pytest -q backend/app/test_auth.py -> 20 passed
D-155	Published requested checkpoint commit/push for current CAPTCHA + Rapid Router UX simplification workspace state to both remotes	2026-02-26	Commit/push executed on main after green npm --prefix frontend run build and targeted python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py
D-154	Executed Rapid Router 10-point UX refactor in 3 phases: compact 5-step header, staged review/submit actions, selection-first cards, reduced action noise, single persistent fix list, helper readability upgrades, admin moved to modal, and UX acceptance targets	2026-02-26	frontend/src/pages/RapidRouter.tsx; npm --prefix frontend run build -> success; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 44 passed, 9 warnings
D-153	Implemented scoped CAPTCHA challenge/verify flow and enforced it on Rapid Router submit + Knowledgebase/POTS message endpoints, with frontend one-time gate UI and token-expiry recovery	2026-02-26	backend/app/main.py, frontend/src/utils/captchaGate.ts, frontend/src/components/CaptchaGateCard.tsx, frontend/src/pages/UnifiedKnowledgebase.tsx, frontend/src/pages/PotsAssistant.tsx, frontend/src/pages/RapidRouter.tsx; python3 -m pytest -q backend/app/test_rapid_router_api_shell.py backend/app/test_knowledgebase_api.py backend/app/test_chat_guidance_api.py backend/app/rapid_router/test_rapid_router_core.py -> 57 passed, 9 warnings; npm --prefix frontend run build -> success
D-152	Produced critical UX/readability audit for Rapid Router and defined a prioritized 10-point simplification game plan	2026-02-26	Audit reviewed key dense zones in frontend/src/pages/RapidRouter.tsx (catalog controls, product cards, right rail, order/options/validation); plan captured in current turn response and tracked as T-067
D-151	Aligned Rapid Router product-card quantity and subtotal controls to a shared bottom baseline using flex anchoring + stabilized placeholder spacing	2026-02-26	frontend/src/pages/RapidRouter.tsx (card h-full flex-column, mt-auto quantity/pricing region, backup plan-code placeholder, shipping-note min-h); cd frontend && npm run build -> success
D-150	Investigated Space boot/wake latency and removed avoidable Rapid Router restart overhead by skipping seed-product rebuild when no seeded-ID backfill is needed	2026-02-25	backend/app/rapid_router/core.py (missing-ID gate around _seed_products()), backend/app/rapid_router/test_rapid_router_core.py (new guard test); python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 20 passed; python3 -m pytest -q backend/app/test_rapid_router_api_shell.py -> 23 passed; startup probes show second-run startup 2882ms vs first-run 6367ms in persisted-storage scenario
D-149	Prepared/published FAQ helper routing fix checkpoint to both required remotes on request	2026-02-25	Publish scope: backend/app/knowledgebase/core.py, backend/app/test_unified_kb_core.py, docs/dev/*, docs/faq/FAQ_ongoing_candidates.csv; pre-publish validation: python3 -m pytest -q backend/app/test_unified_kb_core.py (81 passed), python3 -m pytest -q backend/app/test_knowledgebase_api.py (7 passed, 9 warnings)
D-148	Fixed Rapid Router helper FAQ access so generic concept asks (e.g., network slicing) prioritize FAQ fast-lane with FAQ citation while preserving selected-model compare flows	2026-02-25	backend/app/knowledgebase/core.py (FAQ query context stripping + router-doc FAQ-first branch for RR helper generic asks), backend/app/test_unified_kb_core.py (stronger regression assertions for FAQ route/citation); python3 -m pytest -q backend/app/test_unified_kb_core.py -> 81 passed; python3 -m pytest -q backend/app/test_knowledgebase_api.py -> 7 passed, 9 warnings
D-147	Converted Shipping / Configuration / Payment order-options columns into separate bubble cards for consistent visual grouping	2026-02-25	frontend/src/pages/RapidRouter.tsx (rr-order-options three-column wrappers updated to rounded bordered panels); cd frontend && npm run build -> success
D-146	Normalized Rapid Router card alignment by reserving fixed document and setup-note spacing when optional docs are missing	2026-02-25	frontend/src/pages/RapidRouter.tsx (two fixed doc slots with invisible placeholders; setup-note placeholder block for absent notes); cd frontend && npm run build -> success
D-145	Fixed Rapid Router address-validation suggestion truncation by using full Census matched street line with structured fallback	2026-02-25	backend/app/rapid_router/core.py (_street_from_census_match, validate_us_address mapping), backend/app/rapid_router/test_rapid_router_core.py (2 new regression tests); python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 42 passed, 9 warnings
D-144	Removed duplicate build timestamp display from header (kept single title-area build label)	2026-02-25	frontend/src/components/BrandHeader.tsx (deleted sticky-toolbar build badge); cd frontend && npm run build -> success
D-143	Expanded Rapid Router helper readability (wider rail, larger typography, fuller message/table rendering) while preserving helper logic	2026-02-25	frontend/src/pages/RapidRouter.tsx (grid rail width, helper card/chat sizing, assistant full-width bubbles, larger inline comparison-table preview and CTA); cd frontend && npm run build -> success
D-142	Added PRM lead mode radios (enter_now vs masters_reverse) with conditional validation and mode-aware order outputs	2026-02-25	frontend/src/pages/RapidRouter.tsx (radio controls, conditional PRM input/validation, payload + draft updates), backend/app/rapid_router/core.py (mode normalization/validation + PDF/email label rendering), backend/app/rapid_router/test_rapid_router_core.py (new reverse-PRM success test); python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 17 passed; python3 -m pytest -q backend/app/test_rapid_router_api_shell.py -> 23 passed; cd frontend && npm run build -> success
D-141	Simplified helper comparison-table controls to a single prominent Open table reader CTA for better visibility/selection	2026-02-25	frontend/src/pages/RapidRouter.tsx (HelperMarkdownTable: removed inline expand/copy strip actions, upgraded single CTA styling, retained modal CSV copy); cd frontend && npm run build -> success
D-140	Collapsed Rapid Router catalog search/filter toolbar under a default-closed accordion and kept command-focus compatibility	2026-02-25	frontend/src/pages/RapidRouter.tsx (catalogFiltersOpen state, details/summary wrapper, rapid_router:focus_search auto-open); cd frontend && npm run build -> success
D-139	Fixed Rapid Router helper context-intent regression so generic FAQ/concept asks are not forced into catalog fast-path tables	2026-02-25	backend/app/knowledgebase/core.py (intent detection now uses primary message; selected-context matching remains explicit), backend/app/test_unified_kb_core.py (new regression tests); python3 -m pytest -q backend/app/test_unified_kb_core.py -k \"rapid_router_helper_context\" -> 4 passed; python3 -m pytest -q backend/app/test_unified_kb_core.py -> 81 passed; python3 -m pytest -q backend/app/test_knowledgebase_api.py -> 7 passed
D-138	Triaged restart-time MuPDF FT_New_Memory_Face warning as non-blocking and localized source PDF (atel_re600_manual.pdf)	2026-02-25	Repro command: python3 seed-PDF scan with PyMuPDF over backend/app/rapid_router/seed/assets/*.pdf; warning only on atel_re600_manual.pdf; extraction remained successful (ok pages=5 chars=4261)
D-137	Committed and pushed all outstanding workspace changes to both required remotes on user request	2026-02-25	Modified-set publish including frontend/src/App.tsx, backend/app/rapid_router/seed/assets/atel_w01_u.png, docs/dev/*, docs/faq/FAQ_ongoing_candidates.csv; remotes origin/main, hf-fourtab/main
D-136	Triaged HF env Missing badges and confirmed listed variables are largely optional defaults/presence diagnostics (not immediate runtime failures)	2026-02-25	Code review: frontend/src/components/HealthStatusModal.tsx, backend/app/main.py (/api/health, tab/env defaults), backend/app/router_rag/core.py; guidance delivered with must-set vs optional list
D-135	Set Rapid Router as default landing tab by updating initial tab state, storage-key version, and default flag visibility	2026-02-25	frontend/src/App.tsx; cd frontend && npm run build; cd frontend && npx vitest run --pool=threads --maxWorkers=1 (18 files, 54 tests, all passed)
D-134	Hid Master’s AI and POTS Replacement Q&A toolbox tabs from UI while keeping underlying code paths intact	2026-02-25	frontend/src/App.tsx; cd frontend && npm run build; cd frontend && npx vitest run --pool=threads --maxWorkers=1 (18 files, 54 tests, all passed)
D-133	Replaced incorrect ATEL W01-U seed photo with corrected device image and verified Rapid Router core regression suite	2026-02-25	backend/app/rapid_router/seed/assets/atel_w01_u.png; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 16 passed, 6 warnings
D-132	Prepared and published Rapid Router helper accessibility/table-reader fix bundle checkpoint to both required remotes	2026-02-25	frontend/src/pages/RapidRouter.tsx, docs/dev/*; push targets: origin/main, hf-fourtab/main
D-131	Fixed helper rail accessibility by activating two-column/sticky behavior at lg and ordering rail before main form on single-column layouts	2026-02-25	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build; cd frontend && npx vitest run --pool=threads --maxWorkers=1 (18 files, 54 tests, all passed)
D-130	Added helper table-reader Column focus dropdown with per-column show/hide (first column pinned) for wide comparison-table readability	2026-02-25	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build; cd frontend && npx vitest run --pool=threads --maxWorkers=1 (18 files, 54 tests, all passed)
D-129	Fixed Rapid Router helper comparison-table usability: always-visible Open table reader, stronger inline expand behavior, and sticky first-column/header context	2026-02-25	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build; cd frontend && npx vitest run --pool=threads --maxWorkers=1 (18 files, 54 tests, all passed)
D-128	Reordered Rapid Router right rail (Router helper above Order status) and de-cluttered both cards with shorter copy/chips/actions	2026-02-25	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed
D-127	Committed and pushed startup integrity FAQ/router CSV fix bundle to both required remotes	2026-02-25	commit 914699f; git push origin main -> 13886dc..914699f; git push hf-fourtab main -> 13886dc..914699f
D-126	Published operator runbook for Docker rebuild/redeploy and post-deploy browser cache reset after hashed-asset 404	2026-02-25	Guidance delivered for commit/push to origin + hf-fourtab, wait for HF rebuild complete, then hard refresh/private window to clear stale bundle references
D-125	Fixed startup integrity false alarms in Docker by hardening repo/app path resolution and packaging FAQ corpus in image	2026-02-25	backend/app/knowledgebase/core.py (root/app resolver), Dockerfile (COPY docs/faq /app/docs/faq), backend/app/test_unified_kb_core.py (new resolver tests); python3 -m pytest -q backend/app/test_unified_kb_core.py -> 79 passed; startup integrity probe: faq_entries=551, router_fact_csv_count=3, warnings=[]
D-124	Hardened Auth0 login finalization against silent token timeout and added explicit offline_access scopes	2026-02-25	frontend/src/main.tsx, frontend/src/auth/AuthGate.tsx; cd frontend && npm run build succeeded; cd frontend && npx vitest run --pool=threads --maxWorkers=1 -> 18 passed
D-123	Delivered transfer-oriented one-to-two-page architecture/stack summary for incoming project owner	2026-02-25	Summary prepared from current repo state and ops docs (README.md, backend/app/main.py, docs/dev/open_tasks.md, workflow files)
D-122	Committed and pushed Rapid Router eval25 suite + run-record docs checkpoint to both required remotes	2026-02-25	commit ce1860a; git push origin main -> 7cbce22..ce1860a; git push hf-fourtab main -> 7cbce22..ce1860a
D-121	Diagnosed shard 1-5 failure in Rapid Router eval25 as MSRP-omission semantic miss on W1850 clarify prompt	2026-02-25	`jq '.results[]
D-120	Created Rapid Router-focused 25-case eval suite and executed shard-5 run	2026-02-25	Added docs/evals/unified_kb_eval25_rapid_router_cases.json; cd backend && CHUNK_SIZE=5 START_ID=1 END_ID=25 CASE_TIMEOUT_S=30 OPENAI_MODEL=gpt-5.2 CASES_PATH=../docs/evals/unified_kb_eval25_rapid_router_cases.json OUT_DIR=../docs/evals/shards5_rapidrouter25 TREND_FILE=../docs/evals/shards5_rapidrouter25/unified_kb_eval25_rapidrouter_trend.json ./scripts/run_unified_kb_eval150_chunks.sh -> 24/25, failed IDs [3], avg 23.31ms, p95 30.33ms
D-119	Re-ran full sharded suites on demand and refreshed live metrics for current baseline reporting	2026-02-25	cd backend && CHUNK_SIZE=10 START_ID=1 END_ID=150 OPENAI_MODEL=gpt-5.2 ./scripts/run_unified_kb_eval150_chunks.sh -> 150/150, avg 900.47ms, p95 6316.81ms; cd backend && CHUNK_SIZE=5 START_ID=1 END_ID=75 CASE_TIMEOUT_S=30 OPENAI_MODEL=gpt-5.2 CASES_PATH=../docs/evals/unified_kb_eval75_msrp_verizon_cases.json OUT_DIR=../docs/evals/shards5_eval75 TREND_FILE=../docs/evals/shards5_eval75/unified_kb_eval75_trend.json ./scripts/run_unified_kb_eval150_chunks.sh -> 74/75 (failed IDs [75]), avg 200.59ms, p95 465.47ms
D-118	Re-ran all sharded Unified KB suites (150 + 75) and captured updated aggregate baselines	2026-02-25	cd backend && CHUNK_SIZE=10 START_ID=1 END_ID=150 OPENAI_MODEL=gpt-5.2 ./scripts/run_unified_kb_eval150_chunks.sh -> 150/150; cd backend && CHUNK_SIZE=5 START_ID=1 END_ID=75 CASE_TIMEOUT_S=30 OPENAI_MODEL=gpt-5.2 CASES_PATH=../docs/evals/unified_kb_eval75_msrp_verizon_cases.json OUT_DIR=../docs/evals/shards5_eval75 TREND_FILE=../docs/evals/shards5_eval75/unified_kb_eval75_trend.json ./scripts/run_unified_kb_eval150_chunks.sh -> 74/75 (failed IDs [75])
D-117	Triaged Rapid Router test warnings as non-blocking and captured warning-hygiene follow-up task	2026-02-25	python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 39 passed, 9 warnings; sources: reportlab deprecation + SWIG/PyMuPDF import warnings
D-116	Committed and pushed current CR602 + T-059 + router alias-normalization batch to both required remotes	2026-02-25	commit b87d5d7; git push origin main -> 8d77217..b87d5d7; git push hf-fourtab main -> 8d77217..b87d5d7
D-115	Added deterministic router model alias normalization for hyphen/punctuation variants (MAX-BR1-PRO-5G, XR_60) and regression coverage	2026-02-25	backend/app/knowledgebase/core.py, backend/app/test_unified_kb_core.py; python3 -m pytest -q backend/app/test_unified_kb_core.py -> 77 passed
D-114	Implemented T-059 Rapid Router CSV ingestion validator + dry-run preview/apply admin path with schema/lint checks and duplicate ID/SKU protection	2026-02-25	backend/app/rapid_router/core.py, backend/app/main.py, backend/app/rapid_router/test_rapid_router_core.py, backend/app/test_rapid_router_api_shell.py; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py backend/app/test_rapid_router_api_shell.py -> 39 passed
D-113	Prepared detailed new-thread bootstrap prompt aligned to AGENTS + required dev docs + live working tree state	2026-02-25	Prompt package delivered in chat; continuity anchors: docs/dev/session_handoff.md, docs/dev/decisions.md, docs/dev/open_tasks.md
D-112	Produced ranked 20-item update backlog with complexity/value/risk scoring and identified top 5 implementation targets	2026-02-25	Planning output delivered in chat; promoted top-5 into T-057, T-059, T-060, T-061, T-062
D-111	Added InHand Networks CR602 to Rapid Router seeded catalog with bundled datasheet/manual/image assets and regression assertions	2026-02-25	backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py, backend/app/rapid_router/seed/assets/inhand_cr602.png, backend/app/rapid_router/seed/assets/inhand_cr602_datasheet.pdf, backend/app/rapid_router/seed/assets/inhand_cr602_user_manual.pdf; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 13 passed
D-110	Committed and pushed helper non-store fallback fix checkpoint to both remotes	2026-02-24	commit df60837; git push origin main -> 8f805fb..df60837; git push hf-fourtab main -> 8f805fb..df60837
D-109	Fixed Rapid Router helper model-compare fallback: explicit non-store model asks now bypass store fast path and fall back to standard router-doc compare/spec logic with non-orderable notice	2026-02-24	backend/app/knowledgebase/core.py, backend/app/test_unified_kb_core.py; cd backend && python3 -m pytest -q app/test_unified_kb_core.py app/test_knowledgebase_api.py app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py -> 117 passed
D-108	Committed and pushed T-058 + Rapid Router BoBo/PRM validation hardening checkpoint to both remotes	2026-02-24	commit 7a884c8; git push origin main -> 7215527..7a884c8; git push hf-fourtab main -> 7215527..7a884c8
D-107	Enforced strict PRM Lead format (EL- + exactly 7 digits) with fixed-prefix UI control, backend validation, admin-config validation, and store migration	2026-02-24	frontend/src/pages/RapidRouter.tsx, backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py, backend/app/test_rapid_router_api_shell.py, backend/app/test_tab_final_pass_matrix.py; python3 -m py_compile backend/app/rapid_router/core.py; cd frontend && npm run build; cd backend && python3 -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py app/test_tab_final_pass_matrix.py -> 38 passed
D-106	Added BoBo-conditional required payment fields (Company Name, SPOC, ECPD/VZ Account Number) across Rapid Router UI + submit validation + persisted order/email/PDF outputs	2026-02-24	frontend/src/pages/RapidRouter.tsx, backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py, backend/app/test_tab_final_pass_matrix.py; cd backend && python3 -m pytest -q app/rapid_router/test_rapid_router_core.py app/test_tab_final_pass_matrix.py -> 16 passed; cd backend && python3 -m pytest -q app/test_rapid_router_api_shell.py -> 21 passed; cd frontend && npm run build passed
D-105	Implemented T-058 Rapid Router catalog integration into Unified Knowledgebase (router_docs) with provider injection, deterministic catalog fast paths (list/price/feature/compare), cache-fingerprint wiring, and fallback-to-router-fact behavior	2026-02-24	backend/app/knowledgebase/core.py, backend/app/main.py, backend/app/test_unified_kb_core.py, backend/app/test_knowledgebase_api.py; cd backend && python3 -m pytest -q app/test_unified_kb_core.py app/test_knowledgebase_api.py app/rapid_router/test_rapid_router_core.py -> 92 passed; manual API check via TestClient confirmed retrieval mode deterministic_rapid_router_catalog_list_fast
D-104	Added full-screen comparison-table reader for Rapid Router helper messages (with sticky table headers, improved cell readability, and CSV copy retained)	2026-02-24	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed
D-103	Committed and pushed current Rapid Router + auth stabilization working tree to both remotes	2026-02-24	commit 44c021b; git push origin main and git push hf-fourtab main succeeded
D-102	Updated auth smoke Playwright helper to skip quickly in non-auth local runtime (eliminated false 60s timeouts)	2026-02-24	frontend/e2e/auth.spec.ts; cd frontend && E2E_DISABLE_WEBSERVER=true E2E_BASE_URL=http://127.0.0.1:7860 npx playwright test e2e/auth.spec.ts -> 6 skipped
D-101	Fixed AuthGate refresh-token recovery-flag lifecycle for deterministic re-login/consent recovery behavior	2026-02-24	frontend/src/auth/AuthGate.tsx; cd frontend && npm run build; cd frontend && npx vitest run --pool=threads --maxWorkers=1 -> 18 passed
D-100	Hardened AuthGate timeout env parsing (VITE_AUTH_FINALIZING_WATCHDOG_MS, VITE_AUTH_SILENT_TIMEOUT_MS) against quoted/malformed values	2026-02-24	frontend/src/auth/AuthGate.tsx; cd frontend && npm run build; python3 -m pytest -q backend/app/test_auth.py backend/app/test_rapid_router_api_shell.py backend/app/rapid_router/test_rapid_router_core.py -> 52 passed
D-099	Implemented full Rapid Router UX cleanup bundle (compact order rail, completion chips, jump-to-error links, card/table toggle, shipping indicators, collapsed sections, review modal, helper CSV copy/spacing, session-draft badges, mobile sticky CTA)	2026-02-24	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 11 passed; python3 -m pytest -q backend/app/test_rapid_router_api_shell.py -> 21 passed
D-098	Prepared prioritized UI/UX improvement recommendations for Rapid Router/toolbox cleanup	2026-02-24	Recommendation package delivered in chat; implementation task tracked as T-056
D-097	Added search-driven auto-expand behavior for collapsed Support Toolbox	2026-02-24	frontend/src/App.tsx; cd frontend && npm run build passed
D-096	Collapsed Support Toolbox cards behind a closed-by-default accordion toggle (Open toolbox / Hide toolbox)	2026-02-24	frontend/src/App.tsx; cd frontend && npm run build passed
D-095	Made Ordering assistant and Router selection helper follow together as a single sticky right-column block	2026-02-24	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed
D-094	Updated ground shipping policy to $9.99 with automatic waiver for Standard FWA items; added migration and order/PDF/email shipping breakdown fields	2026-02-24	backend/app/rapid_router/core.py, frontend/src/pages/RapidRouter.tsx; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 11 passed; python3 -m pytest -q backend/app/test_rapid_router_api_shell.py -> 21 passed; cd frontend && npm run build passed
D-093	Updated Peplink MAX BR1 Pro 5G MSRP to $999.00 and added startup migration to correct stale/null persisted MSRP values	2026-02-24	backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 10 passed
D-092	Fixed Router selection helper table rendering by switching assistant bubbles to markdown + added per-table expand/collapse UI	2026-02-24	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed
D-091	Sorted routers by price_primary ascending within each 4G and 5G section on Rapid Router	2026-02-24	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed
D-090	Grouped Rapid Router product catalog into distinct 4G then 5G sections with visual differentiation	2026-02-24	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed
D-089	Committed and pushed Rapid Router reload-reset behavior to both required remotes	2026-02-24	commit a469363; pushed origin/main and hf-fourtab/main
D-088	Changed Rapid Router draft persistence to in-memory only so full reload clears quantities/details while in-app tab switches retain state	2026-02-24	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed
D-087	Verified ATEL RE600 (Black) image was already correct (no-op fix)	2026-02-24	backend/app/rapid_router/seed/assets/atel_re600_black.png; SHA-256 matched Screenshot 2026-02-24 at 11.13.41 AM.png
D-086	Corrected Inseego Wavemaker FX4210 card image (replaced mismatched seed asset with proper Inseego device art)	2026-02-24	backend/app/rapid_router/seed/assets/inseego_wavemaker_fx4210.png; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 9 passed
D-085	Corrected swapped ATEL image assignments: V810AD now uses single-antenna tabletop image and RE600 uses multi-antenna image	2026-02-24	backend/app/rapid_router/seed/assets/atel_v810ad.png, backend/app/rapid_router/seed/assets/atel_re600_black.png; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 9 passed
D-084	Applied ATEL W01-U image hotfix as explicit seed-asset rewrite to corrected user-provided image	2026-02-24	backend/app/rapid_router/seed/assets/atel_w01_u.png; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 9 passed
D-083	Prepared single-commit deployment package for Rapid Router new-device expansion (catalog + assets + migration + tests + template) and executed commit/push workflow	2026-02-24	backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py, backend/app/rapid_router/seed/assets/*, docs/templates/rapid_router_new_devices_upload_template.csv
D-082	Replaced new Rapid Router device photos with exact user-supplied attachment images and enabled startup refresh for existing runtime copies	2026-02-24	backend/app/rapid_router/seed/assets/peplink_b_one_5g.png, backend/app/rapid_router/seed/assets/atel_w01_u.png, backend/app/rapid_router/seed/assets/atel_pw550.png, backend/app/rapid_router/seed/assets/atel_re600_black.png, backend/app/rapid_router/seed/assets/atel_v810ad.png, backend/app/rapid_router/seed/assets/atel_v810vd_bp.png, backend/app/rapid_router/seed/assets/inseego_wavemaker_fx4210.png, backend/app/rapid_router/core.py; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 9 passed
D-081	Added 7 new Rapid Router devices (Peplink B One 5G, ATEL W01-U/PW550/RE600/V810AD/V810VD-BP, Inseego Wavemaker FX4210) with seeded assets and automatic backfill for existing stores	2026-02-24	backend/app/rapid_router/core.py, backend/app/rapid_router/seed/assets/*, backend/app/rapid_router/test_rapid_router_core.py; python3 -m pytest -q backend/app/rapid_router/test_rapid_router_core.py -> 9 passed
D-073	Simplified Rapid Router Ordering Assistant to compact status card with fewer actions and reduced visual complexity	2026-02-24	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed
D-074	Added Rapid Router MSRP support (store schema + product cards + admin add-product MSRP input)	2026-02-24	backend/app/rapid_router/core.py, backend/app/main.py, frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed
D-075	Added workbook-backed required Masters contact dropdown and routed selected contact into order-email To recipients	2026-02-24	backend/app/rapid_router/seed/masters_contacts.xlsx, backend/app/rapid_router/core.py, frontend/src/pages/RapidRouter.tsx; cd backend && python3 -m pytest app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py app/test_tab_final_pass_matrix.py -q -> 29 passed
D-076	Added required Configuration Options with advanced-task validation and per-router pricing rolled into totals/PDF/email	2026-02-24	backend/app/rapid_router/core.py, frontend/src/pages/RapidRouter.tsx; cd backend && python3 -m pytest app/rapid_router/test_rapid_router_core.py app/test_rapid_router_api_shell.py app/test_tab_final_pass_matrix.py -q -> 29 passed
D-077	Committed and pushed Rapid Router MSRP/contact/configuration expansion to both required remotes	2026-02-24	commit 176ff8f; pushed origin/main and hf-fourtab/main
D-078	Remapped Peplink MAX BR1 Pro 5G to use the current MAX BR1 Mini (Wi-Fi) photo and added startup migration for persisted stores	2026-02-24	backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py; cd backend && python3 -m pytest app/rapid_router/test_rapid_router_core.py -q -> 7 passed
D-079	Replaced MAX BR1 Mini (Wi-Fi) product image with requested photo and enforced startup refresh for existing runtime assets	2026-02-24	backend/app/rapid_router/seed/assets/peplink_br1_mini_5g_wifi.png, backend/app/rapid_router/core.py, backend/app/rapid_router/test_rapid_router_core.py; cd backend && python3 -m pytest app/rapid_router/test_rapid_router_core.py -q -> 8 passed
D-080	Added reusable CSV template for Rapid Router new-device intake with MSRP and pricing columns	2026-02-24	docs/templates/rapid_router_new_devices_upload_template.csv
D-072	Fixed mobile overlap where Ordering Assistant covered Router selection helper by making side panel sticky only at lg+ breakpoints	2026-02-24	frontend/src/components/ConversationalSidePanel.tsx; cd frontend && npm run build passed
D-071	Committed and pushed Rapid Router helper chatbot fast-path to both required remotes	2026-02-24	commit 6c6f7dc; pushed origin/main and hf-fourtab/main
D-070	Implemented Rapid Router in-page helper chatbot using existing knowledgebase endpoint in router_docs mode	2026-02-24	frontend/src/pages/RapidRouter.tsx; cd frontend && npm run build passed
D-063	Committed and pushed item 1-5 changes + eval150 rerun results to both required remotes	2026-02-24	commit 54a654c; pushed origin/main and hf-fourtab/main
D-062	Re-ran full unified eval150 (shards10, OpenAI semantic) after implementing items 1-5; achieved zero failures	2026-02-24	docs/evals/shards10/unified_kb_eval150_shards10_summary.json (150/150, 100.0%, failed IDs [])
D-061	Added resilient local /tmp corpus staging in shard runner with manifest fallback generation from chunks	2026-02-24	backend/scripts/run_unified_kb_eval150_chunks.sh (ROUTER_RAG_DATA_DIR=/tmp/router_rag_eval_stage/... confirmed in run logs)
D-060	Added Router RAG fingerprint modes (strict/hybrid/metadata) with timeout-safe hash fallback	2026-02-24	backend/app/router_rag/index.py, backend/app/test_router_rag_module.py (47 passed)
D-059	Added safe .env.codex loading fallback and optional single-process shard execution mode	2026-02-24	backend/scripts/run_unified_kb_eval150_chunks.sh (load_env_file_safe, SINGLE_PROCESS_SHARDS)
D-058	Re-ran unified eval150 in 10-question shards with semantic grading and published updated summary	2026-02-24	docs/evals/shards10/unified_kb_eval150_shards10_summary.json (126/150, 84.0%, failed IDs include 2,3,39-58,116,118)
D-055	Fixed CBA850 token-only routing from weak router-docs path to deterministic lifecycle output	2026-02-20	backend/app/knowledgebase/core.py (_single_lifecycle_only_model_token, auto-mode routing + router-docs bridge), backend/app/test_unified_kb_core.py
D-056	Added regression tests for lifecycle-only single-token routing in router_docs and auto	2026-02-20	cd backend && python3 -m pytest -q app/test_unified_kb_core.py -> 70 passed
D-057	Full backend regression passed after CBA850 routing fix	2026-02-20	cd backend && python3 -m pytest -q -> 316 passed, 9 warnings
D-054	Committed and pushed deep-analysis hardening patch to GitHub and HF four-tab remote	2026-02-20	commit f1e0811; pushed origin/main and hf-fourtab/main
D-051	Patched web fallback timeout budgeting to respect remaining request budget	2026-02-20	backend/app/knowledgebase/core.py (_web_fallback remaining budget guard + timeout cap)
D-052	Hardened parallel index search against stale/shutdown shared executor and added recovery test	2026-02-20	backend/app/knowledgebase/core.py, backend/app/test_unified_kb_core.py
D-053	Deep-analysis verification cycle passed after hardening patches	2026-02-20	cd backend && python3 -m pytest -q -> 314 passed, 9 warnings; cd backend && python3 -m pytest -q app/test_unified_kb_core.py -> 68 passed
D-050	Finalized and pushed enhancement batch commit to both remotes	2026-02-20	commit 925b963; pushed origin/main and hf-fourtab/main
D-044	Implemented targeted fail-ID fixes for masters FAQ clarify over-trigger (102,108) and POTS top-10 objection parsing (63)	2026-02-20	backend/app/knowledgebase/core.py; targeted reruns in docs/evals/shards1_target_102_108/ and docs/evals/shards1_target_75_id63/
D-045	Added stage-budget-exit telemetry and retrieval-mode tracking to eval payloads/summaries	2026-02-20	backend/scripts/unified_kb_eval150.py
D-046	Added runner profile toggle + explicit commit-gate fields (no_new_failed_ids, p95_non_regression) and non-persistent FAQ churn policy by default	2026-02-20	backend/scripts/run_unified_kb_eval150_chunks.sh
D-047	Added regression tests for FAQ medium-confidence bypass and hyphenated top-10 objection handling	2026-02-20	backend/app/test_unified_kb_core.py, backend/app/test_unified_kb_eval150_script.py
D-048	Full backend regression passed after enhancement batch	2026-02-20	cd backend && python3 -m pytest -q -> 312 passed, 9 warnings
D-049	Full OpenAI shard reruns completed (v3)	2026-02-20	docs/evals/shards5_150_balanced_v3/unified_kb_eval150_shards10_summary.json (150/150), docs/evals/shards5_75_balanced_v3/unified_kb_eval150_shards10_summary.json (74/75, fail 3)
D-043	Logged pre-commit low-risk enhancement shortlist and translated into active tasks	2026-02-20	docs/dev/decisions.md, docs/dev/open_tasks.md
D-037	Implemented balanced-profile token/perf caps in router web fallback + POTS synthesis + semantic grader defaults	2026-02-20	backend/app/router_rag/core.py, backend/app/pots_ai/core.py, backend/scripts/unified_kb_eval150.py, backend/scripts/run_unified_kb_eval150_chunks.sh
D-038	Applied OpenAI compatibility fix for POTS completions cap (max_completion_tokens)	2026-02-20	backend/app/pots_ai/core.py
D-039	Clean 150-case rerun (balanced-v2) completed	2026-02-20	docs/evals/shards5_150_balanced_v2/unified_kb_eval150_shards10_summary.json (148/150, fails 102,108)
D-040	Clean 75-case rerun (balanced-v2) completed	2026-02-20	docs/evals/shards5_75_balanced_v2/unified_kb_eval150_shards10_summary.json (74/75, fails 63)
D-041	Full backend regression passed after balanced-v2 changes	2026-02-20	cd backend && python3 -m pytest -q -> 308 passed, 9 warnings
D-042	Before/after comparison package prepared for commit-gate decision	2026-02-20	docs/dev/session_handoff.md, docs/dev/decisions.md
D-019	Full backend deep-dive regression run passed	2026-02-20	python3 -m pytest -q -> 299 passed
D-020	Patched shared bounded retrieval executor path in unified KB	2026-02-20	backend/app/knowledgebase/core.py
D-021	Added runtime health assertion for parallel-search executor flags	2026-02-20	backend/app/test_unified_kb_core.py
D-022	Shard runner now defaults trend and FAQ ongoing paths to OUT_DIR	2026-02-20	backend/scripts/run_unified_kb_eval150_chunks.sh
D-023	Post-patch shard smoke run passed and wrote trend in smoke out-dir	2026-02-20	docs/evals/shards10_deepdive_smoke/unified_kb_eval150_shards10_summary.json
D-024	150-case semantic rerun (shard-5, 30s timeout) completed	2026-02-20	docs/evals/shards5_150_rerun/unified_kb_eval150_shards10_summary.json (146/150)
D-025	75-case MSRP/Verizon semantic rerun (shard-5, 30s timeout) completed	2026-02-20	docs/evals/shards5_75_rerun/unified_kb_eval150_shards10_summary.json (74/75)
D-026	Post-rerun full backend regression passed	2026-02-20	cd backend && python3 -m pytest -q -> 299 passed
D-027	Device comparison table schema updated to user-locked format with hidden evidence column	2026-02-20	backend/app/knowledgebase/core.py
D-028	Comparison schema regression tests added and passing	2026-02-20	cd backend && python3 -m pytest -q app/test_unified_kb_core.py -> 56 passed
D-029	Implemented full guarded 10-suggestion patch set (core + eval tooling + tests)	2026-02-20	backend/app/knowledgebase/core.py, backend/scripts/unified_kb_eval150.py, backend/scripts/run_unified_kb_eval150_chunks.sh, backend/app/test_unified_kb_core.py
D-030	Full backend regression passed after patch set	2026-02-20	cd backend && python3 -m pytest -q -> 308 passed
D-031	150-case shard-5 semantic rerun completed post-patch	2026-02-20	docs/evals/shards10/unified_kb_eval150_shards10_summary.json (144/150, fails 7,86,90,102,108,129)
D-032	75-case MSRP/Verizon shard-5 semantic rerun completed post-patch	2026-02-20	docs/evals/shards5_eval75/unified_kb_eval75_shards5_summary.json (74/75, fails 3)
D-033	Current batch committed and pushed to GitHub + HF four-tab remote	2026-02-20	commit 9e5a3bd; pushed origin/main and hf-fourtab/main
D-034	OpenAI token-usage hotspot analysis completed (no-code step)	2026-02-20	Reviewed backend/scripts/unified_kb_eval150.py, backend/app/pots_ai/core.py, backend/app/router_rag/core.py
D-035	Token-optimization actions ranked by difficulty/perf/token impact and rollout priority	2026-02-20	User-facing ranked matrix prepared; order captured in docs/dev/decisions.md
D-036	Balanced profile recommendation published (performance vs quality)	2026-02-20	Decision logged in docs/dev/decisions.md; task added as T-026

Standard Verification Commands

# Full backend regression
cd backend
python3 -m pytest -q

# Deep-dive runner smoke
cd backend
CHUNK_SIZE=1 START_ID=1 END_ID=1 OUT_DIR=../docs/evals/shards10_deepdive_smoke \
SHARD_WORKERS=1 OPENAI_MODEL=gpt-5.2 ./scripts/run_unified_kb_eval150_chunks.sh