kollguard-mcp

An MCP server for KollGuard. It exposes your SOC 2 / HIPAA security findings and posture as tools, so an AI agent in any MCP client — Claude Code, Cursor, VS Code (Cline/Continue), Windsurf, Zed — can read your live KollGuard data and suggest or apply fixes in your repo.

Read-only by design: it only issues GET requests, authenticated with a read-only kgr_ API key, so it can never change anything in your account.

Setup

1. In KollGuard → Settings → API keys, create a key and copy it (shown once).
2. Add the server to your client's MCP config, with the key:

{
  "mcpServers": {
    "kollguard": {
      "command": "npx",
      "args": ["-y", "kollguard-mcp"],
      "env": { "KOLLGUARD_API_KEY": "kgr_your_key_here" }
    }
  }
}

• Claude Code: .mcp.json in your project root (or claude mcp add).
• Cursor: .cursor/mcp.json.
• VS Code (Cline/Continue) / Windsurf / Zed: their MCP settings use the same shape.

Override the endpoint with KOLLGUARD_API_URL (defaults to https://api.kollguard.com/v1).

Tools

| Tool | What it returns | |---|---| | kollguard_list_findings | Findings (filter by severity / status / framework), each mapped to controls + remediation | | kollguard_get_posture | Open findings, risk score, per-framework gaps, top risks | | kollguard_list_scans | Recent scan runs with severity counts | | kollguard_stats | Headline stats (open findings, by severity/framework, scans last 7d) |

Example prompts

• "Pull my open critical and high KollGuard findings and propose fixes for the repo."
• "What's my SOC 2 posture? Which controls are failing and why?"
• "List database findings and write the SQL migrations to remediate them."