commit-message-quality

Pass

Audited by Gen Agent Trust Hub on Nov 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security issues detected. The skill implements robust defensive measures against common AI agent attack vectors.
  • The instructions explicitly command the agent to treat inspected messages and artifacts as data only and to ignore any instructions embedded within them, which serves as a mitigation against indirect prompt injection.
  • The skill establishes a strict boundary by prohibiting the execution of Git commands, history rewriting, or the execution of external configuration scripts and hooks.
  • It mandates the redaction and exclusion of sensitive content, such as credentials, secrets, and PII, from its output reports regardless of the operational mode.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 21, 2026, 01:08 AM