Feb 19, 2026
skill-scanner
Useful when you want to review skills in any agentic workflow before trusting them in local workflows, CI, or other automated flows.
GitHub repositoryFocus
I use skill-scanner to inspect AI skills, prompts, and instruction files before they enter an agentic workflow. That includes developer tooling, but it is just as relevant anywhere skills are used as trusted setup or execution guidance.
What I care about
The value is catching risky behavior early: secret access, data exfiltration, remote payload pulls, or shell-heavy instructions that do not belong in a trusted skill bundle.
Reference
It combines LLM analysis with VirusTotal and works across multiple agent ecosystems, which makes it practical when the same security review process has to cover more than one toolchain.