AGENT LAB: SKILLS

lucia-auth

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS] (SAFE): The maintenance script scripts/fetch-resources.sh clones documentation from the official lucia-auth and pilcrowOnPaper GitHub repositories. While these are external, they are the authoritative sources for the skill content and the script is intended for developer maintenance rather than runtime execution.
  • [DATA_EXFILTRATION] (SAFE): The templates include network operations to well-known, legitimate endpoints (api.pwnedpasswords.com, googleapis.com, and github.com) for password safety checks and OAuth flows. These operations are implemented correctly and do not exfiltrate sensitive data.
  • [PROMPT_INJECTION] (SAFE): The skill contains instructional guidance for the agent that is aligned with its stated purpose. No behavior overrides or safety bypass attempts were found.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:01 PM