requesting-code-review
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local shell commands using
git. Specifically,SKILL.mdusesgit rev-parseandgit logto identify commit SHAs, whilecode-reviewer.mdusesgit diffto extract code changes for review. These operations are standard for development tools and are necessary for the skill's primary function of code analysis. - [DATA_EXPOSURE_AND_EXFILTRATION]: The skill accesses local repository data (diffs and commit history). There is no evidence of this data being transmitted to external or untrusted domains; the analysis is performed by a subagent within the same context.
- [INDIRECT_PROMPT_INJECTION]: The
code-reviewer.mdtemplate interpolates several placeholders (e.g.,{BASE_SHA},{HEAD_SHA},{PLAN_REFERENCE}) directly into instructions and shell commands. This creates a surface where malicious input in a Git SHA or plan document could theoretically attempt to inject shell commands or instructions. However, as this is a local developer tool, the risk is consistent with standard CLI utilities processing user-defined variables.
Audit Metadata