Skill Audit

•
5
•

Scans AI agent skills, MCP servers, and plugins to detect malicious patterns and security vulnerabilities.

Skill Audit is a static security scanner designed to identify malicious patterns and vulnerabilities across MCP servers, AI agent skills, and plugins. It incorporates over 68 attack signatures categorized across four severity levels, providing comprehensive analysis. The tool integrates seamlessly into various development workflows, offering SARIF output for GitHub Code Scanning, making it an essential solution for maintaining robust security in AI-driven and MCP environments.

Características Principales

01Generates SARIF output for integration with GitHub Code Scanning.
02Offers multiple deployment options: CLI, GitHub Action, multi-arch Docker image, MCP server, Hosted API, and pre-commit hook.
03Detects 68+ attack patterns across four severity levels (Critical, High, Medium, Low).
04Supports scanning content via file path, directory, or directly via content/URL through the hosted API.
05Provides a risk scoring system for identified vulnerabilities.
063 GitHub stars

Casos de Uso

01Performing local security audits of MCP server files and AI agent skills during development.
02Automating security scans in CI/CD pipelines for pull requests to proactively identify vulnerabilities.
03Integrating with AI agents (e.g., Claude Desktop/Cursor) to enable on-demand security assessments.