NVIDIA/SkillSpector
Open-source security scanner for AI agent skills from NVIDIA. Scans skills (from Git repos, URLs, zip files, or directories) for risky patterns including prompt injection, data exfiltration, privilege escalation, malicious code, supply-chain issues, and MCP tool poisoning. Uses static analysis plus optional LLM-based semantic review and can be used in CI/CD or as an MCP server.
