runtz for AI agents

Plug runtz into your favorite AI.

Connect runtz to Claude, Codex, or Gemini and bring security scans directly into the AI workflow you already use.

runtz MCP server

Use the runtz MCP server.

Connect once and give your AI direct access to runtz scans and offline documentation. The MCP server keeps every tool available from the chat, without adding credentials to your prompts.

Set up runtz MCPGet a token

One config block. Every scan.

claude · .mcp.json
{
  "mcpServers": {
    "runtz": {
      "command": "runtz-mcp",
      "env": {
        "RUNTZ_ENDPOINT": "https://engine.runtz.dev",
        "RUNTZ_TOKEN": "rtz_live_..."
      }
    }
  }
}

runtz skills

There's a runtz skill for your AI.

Use Claude, Codex, or Gemini? Add the matching skill and teach your assistant when to scan, which runtz tool to run, and how to return the results.

Browse skills on GitHub

Native to every assistant

Claude

SKILL.md

~/.claude/skills/runtz-security-scans/

A Claude Code & Desktop skill that activates the moment you ask for a security scan.

  • Ships a full CLI reference alongside SKILL.md
  • Routes each request to the right scan automatically
  • Prefers the MCP tools, falls back to the CLI

Give your agent the keys to runtz.

Build the MCP server, add it to your assistant's config, and install the matching skill. The full setup lives in the docs.

MCP setupGet a token