greptile-address
Pass
Audited by Gen Agent Trust Hub on May 7, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes untrusted input from external GitHub PR comments while possessing write access to the repository.
- Ingestion points:
SKILL.md(Steps 3 and 4) instructs the agent to fetch and read Greptile review comments and PR descriptions via the GitHub API. - Boundary markers: The instructions lack explicit boundary markers or directions to treat the comment data as untrusted content.
- Capability inventory: The skill allows the agent to apply code fixes, resolve threads, create commits, and push changes to the branch (
SKILL.mdSteps 6 and 7). - Sanitization: No sanitization or filtering is performed on the incoming review comments before the agent interprets them to determine appropriate fixes.
Audit Metadata