Confidence Rating
Compute a 0 to 100 percent merge-confidence score for a PR after a code review converges, combining application criticality, change complexity, and review-convergence quality. Use after `code-review`, `agent-team-review`, or any review loop completes — to attach a single number plus a per-dimension breakdown to the PR description so reviewers know how much human attention the change warrants. Out…
Code Review
Performs comprehensive code review by analyzing git diffs between the current branch and the default branch (or a specified target). Use when the user requests a code review, diff analysis, PR review, or wants feedback before submitting. Checks project conventions (CODING_GUIDELINES.md, AGENTS.md) first, then evaluates changes across idiomaticity, best practices, clarity, comments & intent, perfo…
Agent Team Review
Parallel multi-agent code review that spawns focused reviewer teammates (correctness, conventions, spec-compliance, contracts, language specialists) via Claude Code Agent Teams, runs CI, cross-references findings between reviewers, and produces a unified merge verdict. Use when the user asks for a team code review, multi-reviewer review, agent team review, parallel diff review across multiple ser…