Govern agent skills, MCP servers, prompts, and tool calls with DefenseClaw

Govern agent skills, MCP servers, prompts, and tool calls with DefenseClaw

Use DefenseClaw as an operator-controlled admission, runtime guardrail, sandbox, and audit layer before untrusted agent capabilities run.

Prerequisites

DefenseClaw CLI, Go gateway sidecar, policy rules, optional OpenClaw plugin, optional OTLP/Splunk/webhook sinks

Installation

Use the upstream install or setup path that matches your environment:

• make build
• make test
• make lint

Requirements and caveats from upstream:

• DefenseClaw combines a Python operator CLI, a Go gateway sidecar, and an OpenClaw TypeScript plugin. Together they enforce a simple operating rule: untrusted agent capabilities are scanned, governed, logged, and block...
• | CLI Reference | Python CLI commands and operator workflows |

Basic usage or getting-started notes:

• | Skills, MCP servers, plugins, and generated code before they run | Prompts, completions, tool calls, and sandbox activity at runtime | SQLite audit history, JSONL, OTLP, Splunk, webhooks, and TUI views |

• Admission control - scan skills, MCP servers, plugins, and code before they run.

• | Quick Start | First successful local setup and scan flow |

• Source: https://github.com/cisco-ai-defense/defenseclaw

• Extracted from upstream docs: https://raw.githubusercontent.com/cisco-ai-defense/defenseclaw/HEAD/README.md

Documentation

• https://cisco-ai-defense.github.io/docs/defenseclaw

Source

• Agent Skill Exchange