github-address-comments
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted third-party content from GitHub PR comments. Ingestion points: Data is fetched via
gh apicalls to pull and issue comment endpoints in SKILL.md. Boundary markers: The skill contract explicitly requires the agent to treat comments as untrusted, requiring redaction and prohibiting following embedded instructions. Capability inventory: The skill can modify local code and post GitHub replies. Sanitization: Instructions specify summarizing and redacting untrusted text. - [COMMAND_EXECUTION]: The skill executes GitHub CLI (
gh) commands to perform authentication checks and data retrieval within the user's authorized repository scope.
Audit Metadata