receiving-code-review
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and act upon code review feedback from external reviewers, which represents untrusted input entering the agent's context.\n
- Ingestion points: Code review feedback provided by human partners or external reviewers (SKILL.md).\n
- Boundary markers: The instructions include a multi-step verification process (READ, UNDERSTAND, VERIFY, EVALUATE) and explicit warnings to avoid blind implementation.\n
- Capability inventory: The skill instructs the use of
grepfor codebase searches andgh apifor interacting with GitHub Pull Request comments.\n - Sanitization: No explicit programmatic sanitization is defined for the input text, relying on the agent's evaluation steps.\n- [COMMAND_EXECUTION]: The skill instructs the agent to use CLI tools such as
grepfor searching the codebase andgh apifor responding to GitHub comments. These are standard developer tools used within the intended scope of the skill.
Audit Metadata