# Aming Claw MCP server

> Gives AI coding agents pushed role, contract, and evidence guidance, and lets them query backlog, graph, and review state via MCP.

- Repository: https://github.com/amingclawdev/aming-claw
- Page: https://hostedmcp.org/servers/amingclawdev/aming-claw
- Transport: —
- Stars: 29

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `AMING_CLAW_BUILD_COMMIT` | no | no |  |
| `GOVERNANCE_PORT` | no | no |  |

## FAQ

### What does Aming Claw add to an AI coding agent?

It pushes the current role, contract, next legal action, required evidence, and verification gate to the agent before it acts, instead of making the agent fetch its own context.

### How do I install Aming Claw for Claude Code?

Add the marketplace, install the plugin, pip install the runtime from the marketplace clone, then start governance and open the dashboard.

### Where is the Aming Claw dashboard served?

It runs locally on port 40000 at the /dashboard path; the root path may return 404.

### What MCP tools does it expose to agents?

Agents can query live project state such as the onboard route guide, backlog, graph, and runtime evidence rather than relying on chat memory.

### Does Aming Claw support parallel workers?

Yes, it offers mf_parallel for one backlog row and mf_batch_parallel for a parent row with child rows; merge, reconcile, and close stay ordered.