refactor-review-gate
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute several shell commands, including
git diffto retrieve the codebase changes andnodeto run internal orchestration and evaluation scripts (e.g.,orchestrate.mjs,guidelines.mjs, andchecklist.mjs). These commands are scoped to the project environment and are necessary for the skill's function as a review gate. - [DATA_EXPOSURE_AND_EXFILTRATION]: While the skill reads the entire
git diffand various internal reports, it does not perform any network operations or external data transmissions. The processing remains local to the execution environment. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input in the form of a
git diff. However, it functions as a read-only reviewer and uses a structured aggregation script (panel-aggregate.mjs) to synthesize findings, which reduces the surface for adversarial data to influence agent behavior.
Audit Metadata