# Skill Intelligence Agent

A RAG-powered agent that answers questions about skills/roles by retrieving
from a vector knowledge base and reasoning over it with tool-calling, exposed
both as a REST API and as an MCP server.

## Architecture
- **RAG layer** (`app/rag`) — Chroma vector store + sentence-transformer embeddings
- **Agent layer** (`app/agent`) — Claude tool-calling loop that queries the RAG layer
- **MCP layer** (`app/mcp`) — exposes the same retrieval tool over MCP for external clients
- **API** (`app/main.py`) — FastAPI endpoints: `/health`, `/ingest`, `/query`
- **Deploy** — Dockerized, deployed to AWS, CI/CD via GitHub Actions

## Local development
```bash
pip install -r requirements.txt
cp .env.example .env  # add your ANTHROPIC_API_KEY
uvicorn app.main:app --reload
```

## Docker
```bash
docker build -t skill-intelligence-agent .
docker run -p 8000:8000 --env-file .env skill-intelligence-agent
```

## Tests
```bash
pytest tests/ -v
```

## Roadmap
- [x] RAG pipeline (ingest + retrieve)
- [x] Agent tool-calling loop
- [x] MCP server exposing retrieval as a tool
- [x] Dockerfile
- [x] CI/CD (test + build + push)
- [ ] AWS deployment (ECS/Lambda)
- [ ] Auth + rate limiting
- [ ] Demo UI / recorded walkthrough