Add local H100 scientist eval tooling

Add 12 scoring & environment improvements with full test coverage

Add model-driven local runtime and dynamic demo flow

Finalize demo flow and training assets

Merge Kush frontend integration, close API 16/UI 10/UI 11

Add API 19 /web fallback route, merge Kush frontend, close UI 07

Add ENV 09 disk persistence, OBS 07/09, TST 11 audit tests, close 10 Max tasks

Add MOD 08 schema tests, V2 training stack, and close MOD 08/JDG 07/API 01/OBS 02

Add JDG 07: reward breakdown logging to CSV and JSONL per episode

Add hybrid Oracle layer and update architecture docs

Recover env judge training stack and sync project tracking

Recover env server client stack and deployment tracking

Add deterministic judge scoring engine (JDG 01-03)

Add AGT 04/05/07 implementations, server integration, and doc updates

Add deterministic alternative suggestion logic for Lab Manager (AGT 06)

Add parse-and-retry loop for Scientist agent with telemetry (AGT 03)

Add per-turn observation formatter for Scientist agent (AGT 02)

Add deterministic protocol validation against scenario constraints (MOD 05)

Type EpisodeState and EpisodeLog with Protocol, ConversationEntry, RewardBreakdown (MOD 04)

Add MOD 11 typed StepInfo/RewardBreakdown and import completed foundation modules

Add future improvements doc, import FND 09 and MOD 01-03 completions

Complete FND 01 and FND 10, update task division with status tracking