skill-vetter

Vetting AI skills for security risks involves verifying origin and author reputation, performing mandatory code review for unsafe patterns, and checking permission scopes for read/write access and network calls before installation.

A security code review for AI agent skills inspects all files for red flags and unsafe patterns, assesses read/write access and network calls, and categorizes risk to produce a structured vetting report for deployment decisions.

Yes, this vetting protocol evaluates sources, code integrity, and permission scopes for AI agent skills shared via ClawdHub, GitHub, or across ecosystems, ensuring safe deployment through reproducible code review and risk classification.

Assess permission scopes and risk classification by checking read/write access, network calls, and command requirements, then categorize risk levels and generate a structured vetting report to guide safe deployment decisions.

A structured vetting report ensures a reproducible vetting process by documenting source verification, code review findings, permission scope assessment, and risk classification, providing clear guidance for safe AI skill deployment decisions.

Never install an AI skill without vetting it first when it requires read/write access, network calls, or command execution, as unsafe patterns and broad permission scopes can introduce significant deployment risks.