Analyze AI agent skills for malicious code and prompt injection risks.
npx skills add https://github.com/Tai-ch0802/skills-bundle --skill skill-vetter-tai-ch0802
Please help me install this Agent Skill. Skill: skill-vetter Source: https://github.com/Tai-ch0802/skills-bundle/tree/main/skill-vetter Command: npx skills add https://github.com/Tai-ch0802/skills-bundle --skill skill-vetter-tai-ch0802
This Skill provides a rigorous security and safety vetting process for AI agent skills, preventing the installation of malicious or risky code.
Use the skill-vetter to vet the skill located at https://github.com/example/new-skill.
High-intent search queries and answers about installing and using this skill.
Security vetting identifies malicious code, prompt injection, and data exfiltration risks by analyzing source reputation, code red flags, and permission scope to classify risk levels and deliver a safety verdict.
Risk assessment for AI agent skills evaluates code for suspicious patterns and unauthorized permissions to assign Low, Medium, High, or Extreme risk classifications that guide safe installation decisions.
Detect prompt injection and data exfiltration risks by executing a security-first code review that scans for suspicious patterns and unauthorized permission requests to generate a safety verdict.
Yes, you can vet an unknown skill from a GitHub repository by running a security audit that analyzes source reputation and permission scope to identify potential malware or unauthorized data access.
Code review for agent security vetting is limited to analyzing source reputation, code patterns, and permission scope, providing a risk classification rather than an absolute guarantee against zero-day vulnerabilities.
AradotsoAnalyze GitHub repositories and software packages for malware indicators and authenticity.
CraftOS-devAnalyze AI agent skill source code for malicious patterns and risk levels.
honysyangDetect permission overreach by comparing declared permissions against actual code behavior.
steven508508Vet AI skills for security risks, permissions, and suspicious code patterns.
