review-loop
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from GitHub Pull Request review threads and CI feedback to drive automated code changes and workflow decisions.
- Ingestion points: GitHub PR review comments, CI check states, and PR descriptions are consumed via the 'gh' CLI and MCP tools.
- Boundary markers: The skill does not implement explicit delimiters or instructions to ignore potential commands embedded in the external review data.
- Capability inventory: The skill possesses significant capabilities, including dispatching sub-agents ('pr-reviewer', 'ci-auto-fix'), invoking secondary skills ('implement-suggestion', 'polish', 'preview-spec'), pushing code to repositories, and editing PR descriptions.
- Sanitization: The skill relies on the downstream safety guardrails of the agents and skills it orchestrates, rather than sanitizing inputs at the orchestration level.
- [COMMAND_EXECUTION]: The skill utilizes the 'gh' (GitHub) and 'git' command-line interfaces to perform repository operations such as querying checks, editing PRs, and pushing local changes. These are standard operations for its stated purpose.
- [EXTERNAL_DOWNLOADS]: The skill interacts with well-known and trusted external services, specifically GitHub and Linear, to manage development workflows and task tracking.
Audit Metadata