← ClaudeAtlas

automated-code-reviewlisted

Systematic review execution engine -- transforms structured human review coaching patterns into autonomous review checklists with pass/fail gates, convention detection, and structured finding production. Use when running autonomous code reviews to ensure systematic coverage and consistent quality.
michaelalber/ai-toolkit · ★ 1 · AI & Automation · score 77
Install: claude install-skill michaelalber/ai-toolkit
# Automated Code Review > "Inspection at the source is called prevention. Inspection after the fact is called sorting." > -- W. Edwards Deming ## Core Philosophy Code review coaching teaches humans how to review; automated code review teaches agents how to execute reviews systematically — the difference between a textbook and a manufacturing process. This skill converts the CACR (Challenge-Attempt-Compare-Reflect) coaching framework into an execution framework: systematic checklists with explicit pass/fail gates that prevent superficial reviews, and convention detection that calibrates findings to the project's own standards rather than abstract ideals. A coaching skill teaches a human to ask "did I check error handling?"; this skill tells the agent exactly HOW — enumerate every fallible call, verify each has an error path, verify the path handles rather than swallows, verify the message carries diagnostic context. It does NOT teach review principles — it assumes those from `code-review-coach` and `security-review-trainer`. It provides the operational framework: checklists, gates, convention detection, and structured output that turn review knowledge into consistent execution. **Non-Negotiable Constraints:** 1. SYSTEMATIC COVERAGE — every category is checked against a concrete checklist; the checklist is the floor, never skipped. 2. CONVENTION CALIBRATION — detect the project's conventions before reporting any style finding; external-standard findings on a differing proj