Antigravity Sub-Agent Bridge (FastMCP)

A containerized FastMCP server that exposes the Google Antigravity CLI (agy --headless) as remote Model Context Protocol (MCP) tools over SSE.

This bridge allows conversational AI clients (such as Google Spark, Claude, Cursor, or ChatGPT) to dispatch full coding and automation sub-agents in a sandboxed, containerized workspace.

---

🚀 Key Features

• Synchronous Sub-Agent Execution (execute_antigravity_sync): Dispatches a coding/automation task directly to agy --headless and returns the output log once the agent finishes (ideal for short 1–3 minute tasks).
• Asynchronous Background Execution (start_antigravity_async): Spawns a long-running Antigravity agent in the background, returning a task_id for decoupled execution.
• Task Polling & Management (get_antigravity_task_status, list_antigravity_tasks): Polls progress, inspects logs, and lists all background sub-agents.
• Workspace Isolation & Sandboxing: Runs within a Docker container mounting only ./workspaces, protecting host system files while giving the agent full build, test, and execution autonomy inside its sandbox.
• Seamless Integration with mcp-server-oauth-proxy: Exposes standard SSE endpoints on port 8080/sse, pluggable directly into Home Assistant's OAuth proxy for secure access from Google Spark.

---

🛠️ Repository Layout

antigravity-mcp-bridge/
├── Dockerfile              # Container image with Python 3.12, Node.js, git, & agy CLI
├── docker-compose.yml      # Service definitions with volume mounts & port bindings
├── requirements.txt        # fastmcp, mcp[cli], pydantic
├── server.py               # FastMCP Server exposing Antigravity agent tools
├── .env.example            # Environment variable template
├── .gitignore              # Ignored artifacts & workspace directories
└── README.md               # Documentation

---

📦 Quickstart & Deployment

1. Configure Environment

Copy the .env.example template:

cp .env.example .env

(Optional: Add your GEMINI_API_KEY or ANTIGRAVITY_API_KEY if required by your agy environment).

2. Start with Docker Compose

docker compose up -d --build

The FastMCP server will start listening on: 👉 http://localhost:8080/sse (or your host LAN IP).

3. Verify Container Health

docker compose logs -f

---

🔗 Connecting to mcp-server-oauth-proxy

To expose this Antigravity sub-agent bridge securely to Google Spark via your Home Assistant OAuth proxy:

1. In Home Assistant, open Settings > Add-ons > MCP Server OAuth Proxy > Configuration.
2. Add a new server entry under servers:

    - client_name: "Google Spark"
      server_name: "Antigravity Bridge"
      server_url: "http://<YOUR_DOCKER_HOST_IP>:8080/sse"
      bearer_token: ""
      client_id: "spark-antigravity"
      client_secret: "secret-antigravity"
      allowed_redirect_uris: "https://oauth-redirect.googleusercontent.com"
      host: "antigravity-oauth.yourdomain.com"
      path_prefix: ""

3. Save and restart the add-on.
4. In Google Spark, add the custom tool pointing to https://antigravity-oauth.yourdomain.com/sse with Client ID spark-antigravity.

---

🧰 Available MCP Tools

1. execute_antigravity_sync

• Description: Executes a task synchronously via agy --headless and waits for output.
• Parameters:
  • prompt (string, required): The goal or coding instruction.
  • sub_path (string, optional): Subdirectory in /workspace to execute within.
  • timeout_seconds (int, default: 300): Max seconds to wait before aborting.

2. start_antigravity_async

• Description: Starts an Antigravity task in the background and returns a task_id.
• Parameters:
  • prompt (string, required): The goal or coding instruction.
  • sub_path (string, optional): Subdirectory in /workspace to execute within.

3. get_antigravity_task_status

• Description: Inspects progress and full logs of a background task.
• Parameters:
  • task_id (string, required): The 8-character ID returned by start_antigravity_async.

4. list_antigravity_tasks

• Description: Lists all active and completed background tasks.