refine-skill-design

Pass

Audited by Gen Agent Trust Hub on May 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a meta-utility for auditing and optimizing AI agent skill definitions.
  • [PROMPT_INJECTION]: No override markers, bypass instructions, or jailbreak patterns were detected.
  • [DATA_EXFILTRATION]: The skill does not perform network operations or access sensitive system files like credentials or SSH keys.
  • [COMMAND_EXECUTION]: The skill instructs the agent to update documentation files locally, which is consistent with its stated purpose of refining skill designs.
Audit Metadata
Risk Level
SAFE
Analyzed
May 7, 2026, 05:42 AM