resolve-pr-comments
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the official GitHub CLI tool (
gh) for both read and write operations, including fetching threads and resolving comments. These commands are appropriately scoped to the skill's functionality and do not exhibit signs of command injection or unauthorized access. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data in the form of PR review comments and bot feedback (Phase 1). It addresses this potential attack vector by requiring the agent to stop and wait for human approval of a generated 'Fix Plan' (Phase 3) before any automated changes are executed by subagents or threads are resolved via the API. This architectural safeguard ensures that a human remains in control of the agent's actions based on external input.
- [DYNAMIC_EXECUTION]: The skill spawns task-specific subagents to perform code fixes (Phase 4). The prompts for these subagents are derived from the triaged PR comments. This use of dynamic task delegation is restricted to the context of the approved fix plan and follows standard agent patterns.
Audit Metadata