DevelopmentPython

agent-governance-toolkit

Kernel-level governance MCP server for AI agents — enforces deterministic policies (tool filtering, budget caps, rate limits, audit logging) instead of prompt-based guardrails. Part of microsoft/agent-lightning (14k★). Run via `npx agentos-mcp-server`.

Free
N/A rating (0 reviews)0 installs824 GitHub stars
agent-frameworkai-agentsai-safetycompliancegovernancemicrosoftowasppolicy-enginepythonsecuritytrustzero-trust

Kernel-level governance MCP server for AI agents — enforces deterministic policies (tool filtering, budget caps, rate limits, audit logging) instead of prompt-based guardrails. Part of microsoft/agent-lightning (14k★). Run via `npx agentos-mcp-server`.

Compatible with

Claude DesktopCursor

Install

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "agent-governance-toolkit": {
      "args": [
        "mcp-server-agent-governance-toolkit"
      ],
      "command": "uvx"
    }
  }
}

Config File Location

Mac: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%\Claude\claude_desktop_config.json

Linux: ~/.config/claude/claude_desktop_config.json

Some servers require additional setup - check the GitHub README for specific instructions.