vett

Vett is a security tool designed to protect artificial intelligence AI agents. It acts like a security guard for the tools or skills that AI agents use. Normally, AI agents might download code from places like GitHub to perform tasks. However, this code could be risky and might steal important information such as passwords or personal data. Vett checks every piece of code before the AI agent uses it to make sure it is safe.

Benefits

Vett helps keep your AI agents secure by scanning their skills. It finds potentially harmful code that could steal your data. This means you can use AI agents with more confidence, knowing that your sensitive information is protected. Vett is also very fast, checking skills in less than 50 milliseconds, and it works automatically without needing complex setup.

Use Cases

This tool is useful for anyone who uses AI agents for their work. Individual developers can easily install Vett to protect their own AI tools. Teams can use Vett to monitor all the AI agents they use, ensuring that only safe skills are being utilized across their organization. It helps prevent AI agents from accessing sensitive company information without permission.

Vibes

Vett has analyzed over 8,000 skills and has identified hundreds of critical security issues. This shows that many skills used by AI agents can indeed pose a risk, and Vett plays a crucial role in identifying these threats.

Additional Information

Vett supports many popular AI agents, including Claude Code, Codex, Cursor, and GitHub Copilot. It uses a secure signing process called Sigstore to guarantee that the scanned code has not been tampered with. For teams, Vett offers features like a private skill registry for internal tools and a clear record of all security checks performed.