review-implementation

# Review Implementation Structured review checklist for verifying implementation completeness. ## Invocation - `/review-implementation` — auto-detect from git diff - `/review-implementation <path>` — review specific files ## Step 1: Detect What Changed ```bash git diff --name-only main # or appropriate base branch ``` Categorize changes: - New files added → component/module review - Transformation/mapping files → rule review - Modified existing files → change review ## Step 2: Structural Checks Run checks appropriate to the project. Read CLAUDE.md for project-specific conventions. ### Universal Checks (all projects) | # | Check | How | |---|-------|-----| | 1 | Source file exists at expected location | Glob | | 2 | Tests exist for new code | Glob for test files | | 3 | Module registered/exported | Grep in registry files | | 4 | No orphaned imports | Grep for unused imports | | 5 | Naming follows project convention | Compare with existing files | ### Project-Specific Checks Read CLAUDE.md for additional checks. Common patterns: - Trait/interface implementations complete - Serialization support (if required) - CLI/dispatch registration - Documentation entries - Example programs registered Build the checklist dynamically from CLAUDE.md conventions + reference implementations. ## Step 3: Build Checks Run project build and test commands (from CLAUDE.md): ```bash # Examples — actual commands from CLAUDE.md make test make lint # or: pytest, cargo test, npm test, etc. ``` Report pass/fail. If tests fail, identify which and suggest fixes. ## Step 4: Semantic Review Read implementation files and assess: 1. **Correctness** — Does the logic match the plan/spec? Edge cases handled? 2. **DRY** — Duplicated logic that should be extracted? Copy-pasted blocks? 3. **KISS** — Unnecessary complexity? Over-engineering? Premature abstraction? 4. **Explainability** — Can every non-trivial change be explained in one sentence? ## Output Format ``` ## Review: [Type] [Name] ### Structural Completeness | # | Check | Status | |---|-------|--------| | 1 | Source file exists | PASS | | 2 | Tests exist | PASS | | ... | ... | ... | ### Build Status - tests: PASS / FAIL - lint: PASS / FAIL ### Semantic Review - Correctness: OK / [issues] - DRY: OK / [duplicated logic in ...] - KISS: OK / [unnecessary complexity in ...] ### Summary - X/Y checks passed - [action items for failures] ``` ## Integration Called automatically at the end of: - `workflow:add-model` (Step 5) - `workflow:add-rule` (Step 6) - `workflow:workflow` (Step 4) Also invocable standalone via `/review-implementation`.