← ClaudeAtlas

deep-reviewlisted

Evidence-based deep code review rubric — severity + confidence findings across maintainability, correctness, security, performance, and testing. Push for ambitious structural simplification (code judo), not just local cleanup. Escalate hands-on triage to the code-reviewer agent.
ulises-jeremias/agent-toolkit · ★ 16 · Code & Development · score 76
Install: claude install-skill ulises-jeremias/agent-toolkit
# Deep Review A strict, evidence-cited code review rubric for when a change needs more than a surface checklist. It pushes the reviewer to be **ambitious about structure** — not merely to identify local cleanup, but to find "code judo" moves: restructurings that preserve behavior while making the implementation dramatically simpler, smaller, and more direct. ## When to use - The user asks for a "deep review", "deep code quality audit", "thermo-nuclear review", "harsh maintainability review", or "review for abstraction quality". - A PR touches large files, shared layers, or risk-prone abstractions. - The `code-reviewer` agent flags that a change needs structural (not stylistic) feedback. ## Review contract Findings carry **severity** and **confidence**, each with an evidence citation (file:line or diff hunk) — no ungrounded opinions: - **Severity:** `critical` (block merge) · `warning` (should fix) · `suggestion` (consider). - **Confidence:** `high` (certain from the evidence) · `medium` (likely, needs a look) · `low` (hypothesis). ## Baseline prompt > Perform a deep code quality audit of the current branch's changes. > Rethink how to structure / implement the changes to meaningfully improve code quality without impacting behavior. > Work to improve abstractions, modularity, reduce spaghetti code, improve succinctness and legibility. > Be ambitious — if there is a clear path to improving the implementation that involves restructuring, pursue it. > Be extremely thorou