What problem does it solve? Automated review bots like PR-Agent and SonarQube generate a mix of genuine bug reports, false positives, and stylistic suggestions that conflict with project conventions. Manually sorting through this feedback, deciding what to fix, and avoiding endless re-review loops consumes significant developer time and risks pushing untested changes. ## Core Features & Use Cases - Feedback Triage Matrix: Fetches PR comments, inline review comments, and CI check status via the GitHub CLI, then categorizes each finding as Accept & Fix, Push Back, or Escalate based on project standards like AGENTS.md, .pr_agent.toml, and sonar-project.properties. - Freshness Verification: Compares the reviewed commit SHA against the current branch HEAD to detect outdated feedback or in-progress reviews before acting on them. - Local Verification & Conditional Push: Runs Nx affected tests and lint checks locally before committing, and requires explicit user approval before pushing to remote. - Use Case: A developer opens a pull request and receives ten PR-Agent suggestions plus three SonarQube findings. The skill inspects all feedback, pushes back on suggestions violating Angular 21 Signal conventions, fixes the two genuine null-safety bugs, verifies locally, and asks before pushing. ## Quick Start Ask the AI to review and address the automated feedback on the current branch's open pull request.