The Agent Skills Directory

[SAFE]: The skill implements a logical security remediation workflow. It correctly prioritizes upgrading vulnerable dependencies over ignoring findings. When an upgrade is not available, it requires the agent to perform a reachability and impact analysis. The skill explicitly forbids autonomous disabling of tools, requiring a human-in-the-loop confirmation (AskUserQuestion) before any significant security posture change. All network operations (WebSearch, WebFetch) are directed toward legitimate research of CVEs and vulnerability databases. No suspicious patterns, obfuscation, or data exfiltration attempts were detected.

fix-security-issue