self-review
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from GitHub pull request descriptions, diffs, and bot-generated review comments.
- Ingestion points: PR descriptions, commit metadata, code diffs, and GitHub review suggestions (SKILL.md).
- Boundary markers: Absent; the skill does not define specific delimiters to separate untrusted data from system instructions.
- Capability inventory: The agent can interact with the GitHub API to post comments and apply code changes (SKILL.md).
- Sanitization: Partially present; the skill instructs the agent to treat review suggestions as hypotheses to be verified and to categorize them (e.g., as 'true regression' vs 'design conflict') before acting, which serves as a conceptual validation layer.
Audit Metadata