Scores AI skills and MCP servers using 14 signals across 4 dimensions, including safety scanning for common attacks.
Sponsored
Mcpskills is a robust trust-scoring and safety scanning platform designed for AI skills and MCP servers. It meticulously evaluates GitHub repositories across four key dimensions – Alive, Legit, Solid, and Usable – utilizing 14 distinct signals to generate a comprehensive trust score. For AI skills and MCP servers, it provides enhanced safety checks to identify critical vulnerabilities like prompt injection, credential theft, and supply chain attacks, based on established attack patterns. Available as an MCP server for direct IDE integration, a user-friendly website, and a powerful API, mcpskills offers a multi-faceted approach to assess and monitor the reliability and security of software components.
主な機能
01Direct IDE integration as an MCP server for Claude Code, Cursor, and other clients
02REST API for programmatic access to scoring, badges, and monitoring
03Repository monitoring to track changes in trust scores over time
04Comprehensive trust scoring across 4 dimensions (Alive, Legit, Solid, Usable)
050 GitHub stars
06Advanced safety scanning for prompt injection, credential theft, and supply chain attacks
ユースケース
01Assessing the trustworthiness and security posture of AI skills and MCP servers before adoption
02Integrating trust score evaluations directly into development workflows and IDEs
03Monitoring GitHub repositories for changes in their reliability, security, and maintenance status
