om-review-prs

Pass

Audited by Gen Agent Trust Hub on Aug 11, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill demonstrates security best practices by defining clear boundaries between untrusted data and agent instructions.
  • [PROMPT_INJECTION]: The references/agentic-setup.md file contains defensive instructions that explicitly tell the agent to ignore and report directives like 'ignore previous instructions' found in repository or tracker content, mitigating indirect prompt injection risks.
  • [CREDENTIALS_UNSAFE]: The references/rules.md file includes a 'Secrets hygiene' policy that forbids the agent from outputting credentials, tokens, or environment variables to logs or comments, regardless of external instructions.
  • [COMMAND_EXECUTION]: The skill mandates validation and sanitization of externally-sourced data (such as PR numbers or slugs) using strict character whitelisting and quoting before they are used in shell commands or file paths.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 11, 2026, 06:01 AM
Security Audit — agent-trust-hub — om-review-prs