confidence-scoring

Pass

Audited by Gen Agent Trust Hub on May 5, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a purely instructional quality-assurance tool and does not contain any executable code, network requests, or sensitive data access.\n- [PROMPT_INJECTION]: No attempts to bypass safety filters or override agent behavior were detected.\n- [DATA_EXFILTRATION]: No patterns for credential harvesting or unauthorized data transmission to external domains are present.\n- [REMOTE_CODE_EXECUTION]: No external script downloads or dynamic code execution patterns were found.\n- [COMMAND_EXECUTION]: The skill references standard development tools and validation commands (like grep, lint, and tests) in an instructional context for project maintenance, which is safe and expected behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
May 5, 2026, 09:05 AM
Security Audit — agent-trust-hub — confidence-scoring