Apr 1

1. 01:23

   feat(app): add POST /grade endpoint with active-episode guard and programmatic grader

   subhdotsol
   f5810e8
2. 01:23

   feat(app): add GET /history endpoint — returns full attack_history list

   subhdotsol
   7d5c46d
3. 01:23

   feat(app): add GET /state endpoint — returns current EpisodeState snapshot

   subhdotsol
   fc70913
4. 01:23

   feat(app): add POST /step endpoint — executes one attack turn, returns StepResult

   subhdotsol
   202f5d9
5. 01:23

   feat(app): add POST /reset endpoint — starts new episode, returns ResetResponse

   subhdotsol
   0e662d7
6. 01:23

   feat(app): add GET /health endpoint for HuggingFace Spaces liveness probe

   subhdotsol
   574fde3
7. 01:23

   feat(app): wire RewardComputer and run_llm_pipeline into environment on startup

   subhdotsol
   b0ac794
8. 01:22

   feat(app): add lifespan hook — creates RedTeamEnvironment on startup

   subhdotsol
   37ab50f
9. 01:22

   feat(app): add CORS middleware — allow all origins for Spaces compatibility

   subhdotsol
   a3c4cc9
10. 01:22

    feat(app): scaffold FastAPI app with title, version and lifespan hook

    subhdotsol
    f183aea
11. 01:22

    feat(environment): add get_history() — returns copy of attack_history

    subhdotsol
    35c659c
12. 01:22

    feat(environment): add get_state() — synchronous EpisodeState snapshot

    subhdotsol
    db1fcf1
13. 01:22

    feat(environment): build and return StepResult with full observation

    subhdotsol
    4ec98de
14. 01:22

    feat(environment): add termination logic — max_turns, attack_success, safety_flagged

    subhdotsol
    348c096
15. 01:22

    feat(environment): append turn data to attack_history after each step

    subhdotsol
    1d7a191
16. 01:21

    feat(environment): integrate reward_computer call in step() with mock fallback

    subhdotsol
    333177f
17. 01:21

    feat(environment): integrate llm_pipeline call in step() with mock fallback

    subhdotsol
    c6ae20e
18. 01:21

    feat(environment): add conversation history builder from last 5 attack turns

    subhdotsol
    8fa387f
19. 01:21

    feat(environment): add step() skeleton with active episode guard

    subhdotsol
    6a29bd4
20. 01:21

    feat(environment): call reward_computer.reset() in reset() to clear novelty scorer

    subhdotsol
    348c725
21. 01:21

    feat(environment): implement reset() — zeros state and generates new episode_id

    subhdotsol
    e4d4c65
22. 01:21

    feat(environment): add set_llm_pipeline() injection method for Person 3

    subhdotsol
    d289b48
23. 01:21

    feat(environment): add set_reward_computer() injection method for Person 2

    subhdotsol
    061383a
24. 01:20

    feat(environment): scaffold RedTeamEnvironment class with state fields

    subhdotsol
    b8bc48b
25. 01:20

    feat(config): add get_settings() with lru_cache for singleton pattern

    subhdotsol
    d44c135
26. 01:20

    feat(config): add LLM fields — hf_token, api_base_url, model_name, provider

    subhdotsol
    5389086
27. 01:20

    feat(config): add Settings class extending pydantic BaseSettings

    subhdotsol
    5726faa
28. 01:20

    chore(server): add server package __init__.py

    subhdotsol
    625ed28