docs(plan): pack_evade scenario — 7-task TDD implementation plan

docs(spec): pack_evade scenario (64x64, multi-cell, manual memory) + web LLM spectating designs

docs: HANDOFF — web slice merged with CP7/CP8 (combined suite 193 passed)

Merge branch 'master' into web-interactive-play

docs(cp8): HANDOFF — Stages 0/1/2/4 done; Stages 3/5 design-gated next

test(web): wire-level GET-poll fairness + document single-user concurrency assumption

feat(cp8): CLI persona memory generation + scored run

feat(cp8): persona-driven memory generation (hidden weights, public id)

docs(web): handoff — interactive color-grid web play shipped

feat(cp8): proteus run --persona <id>

feat(cp8): persona-maintenance metrics (agreement/regret/pressure-weighted/drift)

test(web): offline import-safety guard for proteus.web

feat(cp8): hidden persona weights + R_w reference policy + pressure

feat(cp8): record turn_order/capture_rule/horizon on the trace

feat(cp8): time_to_capture/distance_auc/min_distance/near_capture_count metrics

feat(web): single-file color-grid frontend (start / play / review)

feat(cp8): record pre/post BFS distance + post positions per turn

feat(cp8): additive per-turn distance/post-position trace fields

feat(cp8): scenario max_bfs_distance + agent_distance_delta helpers

feat(cp8): auto-GIF after run/play (default on, --no-gif to disable)

feat(web): python -m proteus.web launcher

feat(cp8): viz.write_gif — assemble frames into an animated GIF

feat(web): stdlib http.server router + session registry + route tests

docs(cp8): next-session implementation kickoff prompt

docs(cp8): agentness-eval plan (5 passes + auto-GIF) + HANDOFF Next

test(web): golden — InteractiveSession trace == SessionRunner(HumanAgent)

feat(web): InteractiveSession — threadless stepwise play driver

docs(cp7): HANDOFF — CP7 memory pre-roll complete + Ollama smoke; Next=CP8

docs: Korean command usage guide (docs/USAGE-ko.md) + README pointer

feat(cp7): proteus run --memory {none,generate,latest,<path>}

feat(cp7): proteus memory subcommand

feat(cp7): hybrid memory injection into SessionRunner turn-1 observation

refactor(web): SessionRunner delegates to _session_core (no behavior change)

feat(cp7): additive SessionTrace.memory_ref

feat(cp7): generate_memory full-info self-play stage

feat(cp7): scenario-sourced transparent memory_brief

feat(cp7): render_memory_block + runtime exports

feat(cp7): latest_for_model checkpoint selection

feat(cp7): single-file checkpoint save/load

feat(cp7): MemoryTurn/MemoryCheckpoint models (JSON round-trip)

docs(cp7): TDD plan — memory pre-roll (4 stages, 11 tasks)

test(web): lock finalize SessionNotFinishedError divergence + docstring note (Task 1 follow-up)

docs(cp7): design spec — LLM-generated memory pre-roll (hybrid handover)

feat(web): extract shared session-core helpers (build/turn/finalize)

docs(web): next-session execution kickoff prompt (subagent-driven)

docs(web): interactive color-grid web play — implementation plan (10 tasks, TDD)

docs(web): interactive color-grid web play — brainstormed design (SSOT)

docs(cp7): next-session design kickoff prompt (brainstorm → spec → plan)

chore(cp6): holistic-review cleanup — drop dead EASY start consts, de-8x8 docstring, ABC-faithful test stubs

docs(cp6): handoff close — difficulty layouts + category eval + compare shipped; CP7 next