Phase 1 engine audit: ENGINE_AUDIT.md + bench-side closures

Quality drive: schema fix, 5 new/revised packs, 4 engine tests, scenario audit

Engine-feature integration: 4 commands + 9 scenario packs + 4 test suites

Add perception ablation grid (observation channel × fog of war)

Phase 1: unified Controller interface for the eval stack

botgen: surface scripted opponents for YAML; drop premature capture_actor

Scenario configs (level+fog per cell) + adversarial-duel duel/interrupts + clearer names

Structured-fog text mode, premium routing, codex descriptions, minimap colour-by-difficulty

Wire bench to vendored training prompt v2 (system/briefing/minimap)

Training-parity minimap (real terrain + legend) + viewer (system/thinking/debrief)

Live-smoke fixes: tool-call wire 400, episode resilience, real PNG minimap

Deterministic scenario-scoped game knowledge + explicit objective

S8: capture_actor tool schema + dispatch; wildcard count 21->22

Eval resilience layer for real OpenRouter runs

cargo tools: enter_transport + unload, 1:1 congruence

Unified Battle Viewer in app.py + run/model playback identity

guard tool: 1:1 congruence with engine Command.guard

Playback: capture model reasoning + per-turn goal tracker + viewer

Scenario-controlled tool allow/deny + default core set

S7 bench: set_stance + patrol tools (schema 1:1, 17==17)

S7 bench: surrender tool + loss outcome (tool schema 1:1, 15==15)

Building & Planning scenario family (user-specified first set)

Bench: consume S9 economy obs + economy win-conditions + full toolset

Add provider-agnostic model agent (vLLM/OpenRouter/Bedrock)