review-with-grok
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (git diffs and pull request content) by passing it to the Grok CLI engine.
- Ingestion points: External diffs are gathered via
git diffandgh pr diffin Step 2. - Boundary markers: The prompt construction in Step 3 includes explicit role and scope definitions to limit the engine's focus to the provided diff.
- Capability inventory: The skill uses the
Bashtool to executegit,gh, andgrok. It explicitly prohibits the automatic application of fixes or execution of suggested commands. - Sanitization: Step 5 implements a mandatory verification process where every finding is checked against the source code in the host session before being reported, effectively mitigating deceptive findings from the engine.
Audit Metadata