skill-vetter

To check AI agent skills for malicious code, run a security vetting process that scans the codebase for red flags, unauthorized access, and data exfiltration risks. This evaluates source origin, author reputation, and permission scopes to ensure safety standards are met.

Permission scope analysis during security vetting determines exactly which files and commands an AI agent skill needs access to. By checking these requested permissions against internal rules, it identifies potential unauthorized access risks before installation.

You can review the source and author reputation of a skill from ClawdHub or GitHub by running a security vetting check. This evaluates where the skill originated, checks author reputation, and analyzes user feedback to assess overall trustworthiness.

The best way to prevent data exfiltration from AI agent skills is performing security-first vetting prior to installation. This process uses internal rules to perform code review and permission analysis, actively checking for malicious patterns and data exfiltration risks.

Yes, security vetting works for skills from any source repository, including ClawdHub and GitHub. It securely assesses trustworthiness and safety by checking for malicious patterns, unauthorized access, and data exfiltration risks using internal rules regardless of the source.