run-reviewable-subtask-loop
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill coordinates local development workflows by executing Git commands (branching, fetching, merging) and project-specific test suites. It incorporates safety logic to verify branch ownership and state, preventing accidental modification of unrelated work.
- [SAFE]: The skill defines rigorous consent boundaries, requiring explicit user authorization for remote operations, CI triggers, and task delegation. It also mandates human review of diffs and plans at multiple stages of the workflow.
- [SAFE]: The skill reads repository documentation and configuration to generate plans. While this is an ingestion surface for external data, the workflow requires sequential human review of all generated plans and code diffs before integration, mitigating indirect prompt injection risks.
Audit Metadata