governance-mcp — Claude Code MCP server: what it does, how to install
An MCP server that pushes AI compliance evidence into Mima and keeps your readiness score current across EU AI Act, ISO 42001, SOC 2 and NIST AI RMF.
claude mcp add governance-mcp -e MIMA_API_KEY=<MIMA_API_KEY> -e MIMA_WORKSPACE_ID=<MIMA_WORKSPACE_ID> -- npx -y @mima-ai/governance-mcpWhat governance-mcp does
The agent gets 10 tools for the AI system registry and the evidence records in a Mima workspace. attest writes a GRC evidence entry; dry_run_attest previews which controls that entry would earn before anything is saved. register_system registers an AI system under EU AI Act Art. 9. For reporting there's get_posture with the overall score and a per-framework split, plus check_gates returning gate status with exit codes and suggest_gates for prioritised recommendations. One attestation maps onto all four frameworks at once, so the same record counts everywhere.
Setup is an npx -y @mima-ai/governance-mcp command in the MCP config of your editor or CLI. Two values are required first: MIMA_API_KEY and MIMA_WORKSPACE_ID, both from the Mima dashboard. Sign up there and copy the key before you wire anything in — without them the server has nothing to authenticate with.
If attestations belong in app code rather than agent chat, there's a Python SDK: pip install mima-governance. Teams that must show EU AI Act evidence on every step will get more from check_gates, because exit codes plug straight into a pipeline; hand-collecting records each time is slower and easy to get wrong.
What it helps with
- log an AI-risk evidence record in one call
- preview which controls an attestation earns before writing it
- register an AI system under EU AI Act Art. 9
- compare readiness scores across four frameworks
- get gate pass/fail with exit codes for a pipeline
Example prompts for the agent
- Check our compliance posture across all four frameworks.
- Dry-run an AI risk assessment for loan-scorer and show the mapped controls.
How to install governance-mcp
Platform CLI
claude mcp add governance-mcp -e MIMA_API_KEY=<MIMA_API_KEY> -e MIMA_WORKSPACE_ID=<MIMA_WORKSPACE_ID> -- npx -y @mima-ai/governance-mcpReplace <…> with your own values
Platform CLI
codex mcp add governance-mcp --env MIMA_API_KEY=<MIMA_API_KEY> --env MIMA_WORKSPACE_ID=<MIMA_WORKSPACE_ID> -- npx -y @mima-ai/governance-mcpReplace <…> with your own values
Config file
{
"mcpServers": {
"governance-mcp": {
"command": "npx",
"args": [
"-y",
"@mima-ai/governance-mcp"
],
"env": {
"MIMA_API_KEY": "<MIMA_API_KEY>",
"MIMA_WORKSPACE_ID": "<MIMA_WORKSPACE_ID>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"governance-mcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@mima-ai/governance-mcp"
],
"env": {
"MIMA_API_KEY": "<MIMA_API_KEY>",
"MIMA_WORKSPACE_ID": "<MIMA_WORKSPACE_ID>"
}
}
}
}Replace <…> with your own values
Platform CLI
gemini mcp add -e MIMA_API_KEY=<MIMA_API_KEY> -e MIMA_WORKSPACE_ID=<MIMA_WORKSPACE_ID> governance-mcp npx -y @mima-ai/governance-mcpReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about governance-mcp
How do I install governance-mcp in Claude Code?
Run claude mcp add governance-mcp -e MIMA_API_KEY=<MIMA_API_KEY> -e MIMA_WORKSPACE_ID=<MIMA_WORKSPACE_ID> -- npx -y @mima-ai/governance-mcp. Grab both values from the Mima dashboard first.
What does governance-mcp do?
It sends compliance evidence to your Mima workspace and refreshes readiness scoring for EU AI Act, ISO 42001, SOC 2 and NIST AI RMF. A single record lands against all four frameworks.
Can I test an attestation without saving it?
Yes — dry_run_attest lists the controls a record would earn and writes nothing. Handy for agreeing on wording before the evidence reaches the registry.
