Multi-agent Claude Code plugin: parallel specialized reviewers (bugs, security, CLAUDE.md compliance, git history, test coverage), confidence scoring, terminal or inline GitHub PR reporting.
Skill Claude CodeCodex
Shared rubric for scoring code-review findings 0-100 and rejecting false positives. Every reviewer agent in the agentic-code-review plugin cites this before emitting a finding, and the orchestrator applies the same threshold when filtering. Use when scoring a review finding or deciding whether an issue is real signal.
Skill Claude CodeCodex
How the agentic-code-review orchestrator renders findings — the terminal report format and the GitHub inline-comment mechanics (gh api calls, full-SHA anchoring, suggestion blocks, one-comment-per-issue dedupe). Use when formatting a code-review result for the terminal or posting review comments to a pull request.