docs: update HANDOFF architecture + paths for game/web layout

docs(pack_evade): record real-LLM smoke (ollama gpt-oss:120b-cloud — run + spectate, passed)

Merge branch 'web-spectate'

feat(pack_evade): persona weight-vector memory generation (drop manual; reference policy self-play per agentness design)

docs(pack_evade): handoff — 64x64 multi-cell evasion scenario shipped

docs(spectate): handoff — watch an LLM play in color shipped

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

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

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

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

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

docs(cp5): note stdin-EOF fix + empty-turns terminal-flag deferral (103 passed)

docs(cp5): handoff close — human play + trace viz shipped; CP6 next

docs(cp4): handoff update — CLI + cloud providers + token accounting shipped; CP5 kickoff prompt; gitignore .env

docs(cp3): handoff record + CP4 pointer