merge-request-review
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting and analyzing untrusted external content, which is an inherent part of its code-review function.
- Ingestion points: Processes PR/MR descriptions, linked issues, repository instructions, and code diffs as specified in the Workflow section of
SKILL.md. - Boundary markers: The skill does not define explicit delimiters to isolate untrusted data from the agent's instructions, relying instead on the agent's ability to distinguish intent.
- Capability inventory: Utilizes read-oriented operations such as
git diff, platform API calls, and isolated worktrees, with instructions to avoid destructive operations like overwriting user changes. - Sanitization: No specific sanitization or filtering of external input is mentioned; however, the skill prioritizes checking tests and stated intent before production code, which serves as a logical guardrail.
Audit Metadata