perception

Pass

Audited by Gen Agent Trust Hub on Jun 27, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill provides cognitive instructions for detail-oriented tasks like contract review and code analysis without introducing new executable capabilities.
  • [NO_CODE]: No external scripts, dependencies, or binary files are referenced or included in the skill definition.
  • [PROMPT_INJECTION]: The instructions are benign and focused on task execution, with no attempts to bypass safety filters or override system rules.
  • [DATA_EXFILTRATION]: The skill does not perform network operations or access sensitive system paths, ensuring no risk of data exposure.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 27, 2026, 07:47 PM
Security Audit — agent-trust-hub — perception