docs: add initial README with project overview and HuggingFace Spaces frontmatter

README.md

@@ -0,0 +1,74 @@
+---
+title: RedTeamOS
+emoji: 🛡️
+colorFrom: red
+colorTo: purple
+sdk: docker
+pinned: false
+license: mit
+---
+
+# RedTeamOS
+
+AI Red-Teaming Environment for Safety Research.  
+Built for the Meta PyTorch OpenEnv Hackathon.
+
+## Quickstart
+
+```bash
+cp .env.example .env
+# fill in HF_TOKEN, ANTHROPIC_API_KEY
+
+pip install -r requirements.txt
+uvicorn server.app:app --reload --port 8000
+```
+
+## API Endpoints
+
+| Method | Endpoint   | Description                        |
+|--------|------------|------------------------------------|
+| GET    | /health    | Liveness probe                     |
+| POST   | /reset     | Start a new episode                |
+| POST   | /step      | Execute one attack turn            |
+| GET    | /state     | Get current episode state          |
+| GET    | /history   | Get full attack history            |
+| POST   | /grade     | Grade a completed episode          |
+
+## Example Usage
+
+```python
+import httpx, asyncio
+
+async def run():
+    async with httpx.AsyncClient(base_url="https://your-space.hf.space") as client:
+        # Start episode
+        reset = await client.post("/reset")
+        print(reset.json())
+
+        # Attack step
+        action = {
+            "strategy_type":   "roleplay",
+            "target_category": "privacy",
+            "intensity":       0.5,
+            "framing":         "Your attack framing here..."
+        }
+        step = await client.post("/step", json=action)
+        print(step.json())
+
+asyncio.run(run())
+```
+
+## Task Difficulties
+
+| Task   | Max Turns | Strategies Available |
+|--------|-----------|----------------------|
+| easy   | 5         | roleplay, hypothetical |
+| medium | 8         | + persona_switch, authority_claim |
+| hard   | 10        | all strategies |
+
+## Docker
+
+```bash
+docker build -t redteam-env .
+docker run -p 7860:7860 --env-file .env redteam-env
+```