address-code-review
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources that could contain malicious instructions designed to manipulate the agent.
- Ingestion points: External data is fetched from GitHub Pull Request review threads and comments via
gh apiand GraphQL, and parsed from markdown documents within the repository. - Boundary markers: The instructions do not define explicit delimiters or warnings to the agent to ignore instructions embedded within the feedback.
- Capability inventory: The agent is authorized to use
gh apito post replies and resolve threads, and has file-write capabilities to apply agreed-upon changes. - Sanitization: The skill implements a robust human-in-the-loop process, mandating that the agent must present each comment to the user and wait for a decision before recording an action or applying changes. This serves as a primary manual sanitization step.
Audit Metadata