Agent.md

Deployment Manager Agent.md

This file serves as a guide for further agents regarding deployment best practices and tips for Automaker on Hugging Face Spaces.

1. Deployment Configuration

The app is configured to run as a Docker container on port 7860.

Target Spaces

• Target: Leon4gr45/hyp
• Frontend Port: 7860

Mandatory Endpoints

• /health: Returns HTTP 200 when ready.
• /api-docs: Documents all API endpoints.

2. API Documentation

/health

• Method: GET
• Purpose: Check server health
• Response:

  {
    "status": "ok",
    "timestamp": "ISO Date String",
    "version": "1.0.0"
  }
  

/api-docs

• Method: GET
• Purpose: Document all available API endpoints

Functional Endpoints (Example)

• Method: POST
• Path: /api/agent/chat
• Purpose: Send message to AI agent

3. Deployment Workflow

To redeploy, use:

# Deploy to Leon4gr45/hyp
hf upload Leon4gr45/hyp . --repo-type=space --token <YOUR_TOKEN>

OpenCode CLI Authentication

The Space is configured to automatically log in to OpenCode if the OPENCODE_API_KEY or JULES_API_KEY environment variables are set in the Space secrets.

GitHub CLI Authentication

To enable GitHub operations (like PR creation), provide a GitHub Personal Access Token (PAT) via the GITHUB_API_TOKEN environment variable in the Space secrets.

Monitor logs via:

• Build logs: curl -N -H "Authorization: Bearer <YOUR_TOKEN>" "https://huggingface.co/api/spaces/Leon4gr45/hyp/logs/build"
• Run logs: curl -N -H "Authorization: Bearer <YOUR_TOKEN>" "https://huggingface.co/api/spaces/Leon4gr45/hyp/logs/run"

Tips

• Ensure the Dockerfile builds both the UI and Server.
• The server must serve the UI static files and provide a catch-all route for SPA.
• Data is stored in /app/data inside the container.
• The repository is cloned from https://github.com/JsonLord/automaker.git in the Dockerfile.