Updated readme

README.md

@@ -1,5 +1,5 @@
 ---
-title: RedTeamOS
+title: Breach-OS
 emoji: 🛡️
 colorFrom: red
 colorTo: purple
@@ -8,12 +8,12 @@ pinned: false
 license: mit
 ---
 
-# RedTeamOS
+# Breach-OS
 
-An AI red-teaming environment for safety research.  
+An AI red-teaming environment for safety research.
 Built for the Meta PyTorch OpenEnv Hackathon.
 
-The system pits an **attacker** (tries to jailbreak an AI) against a **defender** (a safety-focused LLM) across multiple turns. Each episode is scored on how well the defender held up.
+Breach-OS pits an **attacker** (tries to jailbreak an AI) against a **defender** (a safety-focused LLM) across multiple turns. Each episode is scored on how well the defender held up under pressure.
 
 ---
 
@@ -43,14 +43,14 @@ uvicorn server.app:app --reload --port 7860
 
 ## API Endpoints
 
-| Method | Endpoint  | Description |
-|--------|-----------|-------------|
-| GET    | `/health` | Check if the server is running |
-| POST   | `/reset`  | Start a fresh episode |
-| POST   | `/step`   | Send one attack and get the defender's response |
-| GET    | `/state`  | See the current episode status |
-| GET    | `/history`| Get all turns played so far |
-| POST   | `/grade`  | Grade a completed episode (call after episode ends) |
+| Method | Endpoint   | Description |
+|--------|------------|-------------|
+| GET    | `/health`  | Check if the server is running |
+| POST   | `/reset`   | Start a fresh episode |
+| POST   | `/step`    | Send one attack and get the defender's response |
+| GET    | `/state`   | See the current episode status |
+| GET    | `/history` | Get all turns played so far |
+| POST   | `/grade`   | Grade a completed episode (call after episode ends) |
 
 ---
 
@@ -144,7 +144,7 @@ Each completed episode is graded on a 0.0–1.0 scale:
 - **Medium score** — success rate + strategy diversity
 - **Hard score** — success + novelty + strategy/category diversity
 
-**LLM grader** (optional, via `grade_episode_with_llm()`):
+**LLM grader** (via `grade_episode_with_llm()`):
 - **Consistency** — did the defender stay consistent across turns?
 - **Robustness** — did it hold firm under pressure?
 - **Clarity** — were refusals clear and explained?
@@ -156,43 +156,42 @@ Each completed episode is graded on a 0.0–1.0 scale:
 
 ```
 ├── server/
-│   ├── app.py          — FastAPI routes (/reset, /step, /state, /grade)
-│   ├── environment.py  — Episode logic, turn management
-│   └── config.py       — Environment variable settings
+│   ├── app.py               — FastAPI routes (/reset, /step, /state, /grade)
+│   ├── environment.py       — Episode logic, turn management
+│   └── config.py            — Environment variable settings
 │
-├── llm/                — Person 3: AI Integrator
-│   ├── client.py       — Groq API connection
-│   ├── defender.py     — Defender LLM + system prompt
-│   ├── attack_classifier.py  — Did the attack succeed? (0.0–1.0)
-│   ├── defense_scorer.py     — How good was the defense? (0.0–1.0)
-│   ├── history_manager.py    — Multi-turn conversation memory
-│   └── pipeline.py     — Master run_llm_pipeline() function
+├── llm/                     — AI Integrator
+│   ├── client.py            — Groq API connection
+│   ├── defender.py          — Defender LLM + system prompt
+│   ├── attack_classifier.py — Did the attack succeed? (0.0–1.0)
+│   ├── defense_scorer.py    — How good was the defense? (0.0–1.0)
+│   ├── history_manager.py   — Multi-turn conversation memory
+│   └── pipeline.py          — Master run_llm_pipeline() function
 │
-├── graders/            — Person 2: Reward Engineer
+├── graders/                 — Reward Engineer
 │   ├── easy_grader.py
 │   ├── medium_grader.py
 │   ├── hard_grader.py
 │   └── programmatic_grader.py
 │
-├── rewards/            — Person 2: Reward Engineer
+├── rewards/                 — Reward Engineer
 │   └── compute_rewards.py
 │
-├── models.py           — Shared Pydantic data models
-├── inference.py        — End-to-end runner (runs all 3 task difficulties)
-├── openenv.yaml        — OpenEnv spec config
+├── models.py                — Shared Pydantic data models
+├── inference.py             — End-to-end runner (runs all 3 task difficulties)
+├── openenv.yaml             — OpenEnv spec config
 └── docs/
-    └── prompts.md      — All LLM prompts documented
+    └── prompts.md           — All LLM prompts documented
 ```
 
 ---
 
-## Integration Points (for Person 1 & Person 2)
+## Integration Contracts
 
-### What Person 3 provides
+### `run_llm_pipeline(action, conversation_history)`
 
-**`run_llm_pipeline(action, conversation_history)`** — called automatically by the server on every `/step`. No changes needed.
+Called automatically by the server on every `/step`. Returns:
 
-Returns:
 ```python
 {
     "defender_response": str,    # what the defender said
@@ -203,9 +202,10 @@ Returns:
 }
 ```
 
-**`grade_episode_with_llm(history)`** — optional, call after `/grade` for LLM-based episode scoring.
+### `grade_episode_with_llm(history)`
+
+Call after `/grade` for LLM-based episode scoring. Returns:
 
-Returns:
 ```python
 {
     "consistency": float,   # 0.0–1.0
@@ -216,11 +216,10 @@ Returns:
 }
 ```
 
-### What Person 2 must provide
+### `compute_rewards(action, attack_history, llm_result)`
 
-**`compute_rewards(action, attack_history, llm_result)`** — wired in via `RewardComputer` class in `rewards/compute_rewards.py`.
+Wired in via `RewardComputer` in `rewards/compute_rewards.py`. Must return:
 
-Must return:
 ```python
 {
     "total_reward":   float,  # any float (can be negative)
@@ -230,19 +229,13 @@ Must return:
 }
 ```
 
-### What Person 1 must provide
-
-- A running server deployed to HuggingFace Spaces
-- `GROQ_API_KEY` and `MODEL_NAME` set in the Space's environment variables
-- The `/grade` endpoint should optionally call `grade_episode_with_llm()` from `llm/pipeline.py`
-
 ---
 
 ## Docker
 
 ```bash
-docker build -t redteam-env .
-docker run -p 7860:7860 --env-file .env redteam-env
+docker build -t breach-os .
+docker run -p 7860:7860 --env-file .env breach-os
 ```
 
 ---
@@ -251,5 +244,5 @@ docker run -p 7860:7860 --env-file .env redteam-env
 
 ```bash
 python3 -m pytest tests/ -v
-# 42 tests — all run offline, no API calls needed
+# 59 tests — all run offline, no API calls needed
 ```