# kakunin-ai/kakunin-mcp [Health: Active]

**Category:** 🔒 Security  
**Repository:** https://github.com/kakunin-ai/kakunin-mcp  
**GitHub Stars:** 1  
**npm Downloads (last month):** 313  
**Views:** 4  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/kakunin-ai-kakunin-mcp

## Description
Compliance and identity for AI agents — verify an agent's certificate scope, read its behavioral risk score, and append to an immutable audit trail. X.509 identity issued via AWS KMS; MiCA / EU AI Act aligned. npx -y @kakunin/mcp

## Tools
Capabilities this server exposes over MCP:

- **verify_agent_scope** — Check whether this agent is authorised to perform an action before executing it. Verifies the active X.509 certificate, permitted_actions scope, financial limits, and revocation status.
- **check_risk_score** — Retrieve the agent's rolling 30-day risk score, band (`low`/`medium`/`high`), drift trend, and actionable guidance. No input required.
- **audit_log_append** — Append a behavioral event to the agent's immutable audit log. Returns risk score + transaction ID. Events scoring ≥ 0.85 auto-trigger a certificate revocation check.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "kakunin-mcp": {
    "command": "npx",
    "args": ["-y","@kakunin/mcp"],
    "env": {
      "KAKUNIN_API_KEY": "",
      "KAKUNIN_AGENT_ID": ""
    }
  }
}
```

**Requires environment variables:** `KAKUNIN_API_KEY`, `KAKUNIN_AGENT_ID` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What kakunin-ai/kakunin-mcp MCP server does

The kakunin-ai/kakunin-mcp MCP server exposes Kakunin's AI agent compliance API through Model Context Protocol. It is intended for agents that need to establish whether a requested operation is within their assigned authority, inspect their current behavioral risk, and record events associated with their activity.

Authorization checks use the active X.509 certificate and evaluate the agent's permitted action scope, financial limits, and certificate revocation status. The service is described as using X.509 identity issued through AWS KMS and as being aligned with MiCA and the EU AI Act. Those compliance references describe the service's intended context; they do not replace an application's own legal or operational review.

## How it works

The kakunin-ai/kakunin-mcp MCP server acts on behalf of one Kakunin agent, identified by an agent ID. A Kakunin API key authenticates requests, while the server uses the configured base URL to reach the API. The default endpoint is `https://kakunin.ai`, with an optional override for other environments.

`verify_agent_scope` accepts an action description, venue, and USD amount. It returns whether the action is allowed, along with a reason, certificate status, and permitted actions. `check_risk_score` requires no input and returns the rolling 30-day score, a low/medium/high band, the drift trend, and guidance. `audit_log_append` records a behavioral event and returns a risk score and transaction ID. Events with a score of at least 0.85 trigger a certificate revocation check.

## Setup and configuration

The kakunin-ai/kakunin-mcp MCP server requires Node.js 18 or newer, a Kakunin API key, and an agent ID obtained through the Kakunin dashboard. Configure it as a local stdio server with:

```json
{
  "command": "npx",
  "args": ["-y", "@kakunin/mcp"],
  "env": {
    "KAKUNIN_API_KEY": "kak_live_...",
    "KAKUNIN_AGENT_ID": "agt_..."
  }
}
```

`KAKUNIN_API_KEY` accepts live or test keys, and `KAKUNIN_AGENT_ID` selects the agent represented by the server. `KAKUNIN_BASE_URL` is optional and changes the API base URL. Test keys use Kakunin's sandbox CA, have no cost, and include 100 free certificates per day according to the provided setup notes.

## Tools and capabilities

- Verify whether a proposed action fits the certificate and permitted scope.
- Check financial limits and certificate revocation status before execution.
- Read a 30-day risk score, risk band, trend, and recommended guidance.
- Append structured behavioral events to an immutable audit log.
- Receive a transaction ID and risk result after an audit append.

The README identifies Claude and Cursor as supported usage contexts, as well as other MCP-compatible runtimes. The server does not itself execute the proposed business action; `verify_agent_scope` supplies an authorization result for the calling workflow to use.

_Full upstream README: https://allmcps.com/mcp/kakunin-ai-kakunin-mcp/readme_