resolve-pr-parallel

This skill is plausibly legitimate for automating PR comment resolution, but it carries moderate-to-high supply-chain and operational risk because it delegates code changes, commits, pushes, and GitHub GraphQL mutations to autonomous subagents. The primary concerns are autonomy abuse (automated commits/pushes and resolving threads without per-action human approval) and the broad shell permissions granted to run git/gh commands (which implicitly use user credentials). There is no evidence of direct malicious payloads, remote download-execute chains, or hardcoded exfiltration endpoints in the provided fragment, so confirmed malware is unlikely. However, because the capability can modify and publish repository code and use repository credentials, treat this skill as risky: require explicit human review, tight scope/approval controls, audit logging, and least-privilege credentials before use.