skill-vetter

To vet an AI skill for security vulnerabilities, you can use an automated vetting tool that checks for unauthorized data exfiltration, malicious code, and permission misalignment. It categorizes the risk level to guide your installation decision.

Permission scope evaluation in AI skill vetting ensures that a skill's requested permissions align strictly with its intended functionality. This prevents over-privileged access and reduces potential security risks during installation.

You can perform a code review on an untrusted AI skill from GitHub by running a vetting tool with the repository path. This provides a step-by-step protocol to assess code quality and identify potential security threats.

Yes, you can classify the risk level of an AI skill automatically using a vetting tool. It categorizes the threat level by combining automated security checks with manual review steps to guide your installation decision.

Yes, AI skill vetting works without external dependencies. The vetting process relies on internal scripts to execute automated checks and guide manual code review steps, ensuring a thorough assessment independently.

The limitation of automated vetting for AI skills is that automated checks alone may miss complex logical flaws. The process requires manual review steps to thoroughly assess code quality and ensure permission scopes align with intended functionality.