$npx -y skills add jackspace/ClaudeSkillz --skill code-review_mrgoonieUse when receiving code review feedback (especially if unclear or technically questionable), when completing tasks or major features requiring review before proceeding, or before making any completion/success claims. Covers three practices - receiving feedback with technical rigo
| 1 | # Code Review |
| 2 | |
| 3 | Guide proper code review practices emphasizing technical rigor, evidence-based claims, and verification over performative responses. |
| 4 | |
| 5 | ## Overview |
| 6 | |
| 7 | Code review requires three distinct practices: |
| 8 | |
| 9 | 1. **Receiving feedback** - Technical evaluation over performative agreement |
| 10 | 2. **Requesting reviews** - Systematic review via code-reviewer subagent |
| 11 | 3. **Verification gates** - Evidence before any completion claims |
| 12 | |
| 13 | Each practice has specific triggers and protocols detailed in reference files. |
| 14 | |
| 15 | ## Core Principle |
| 16 | |
| 17 | **Technical correctness over social comfort.** Verify before implementing. Ask before assuming. Evidence before claims. |
| 18 | |
| 19 | ## When to Use This Skill |
| 20 | |
| 21 | ### Receiving Feedback |
| 22 | Trigger when: |
| 23 | - Receiving code review comments from any source |
| 24 | - Feedback seems unclear or technically questionable |
| 25 | - Multiple review items need prioritization |
| 26 | - External reviewer lacks full context |
| 27 | - Suggestion conflicts with existing decisions |
| 28 | |
| 29 | **Reference:** `references/code-review-reception.md` |
| 30 | |
| 31 | ### Requesting Review |
| 32 | Trigger when: |
| 33 | - Completing tasks in subagent-driven development (after EACH task) |
| 34 | - Finishing major features or refactors |
| 35 | - Before merging to main branch |
| 36 | - Stuck and need fresh perspective |
| 37 | - After fixing complex bugs |
| 38 | |
| 39 | **Reference:** `references/requesting-code-review.md` |
| 40 | |
| 41 | ### Verification Gates |
| 42 | Trigger when: |
| 43 | - About to claim tests pass, build succeeds, or work is complete |
| 44 | - Before committing, pushing, or creating PRs |
| 45 | - Moving to next task |
| 46 | - Any statement suggesting success/completion |
| 47 | - Expressing satisfaction with work |
| 48 | |
| 49 | **Reference:** `references/verification-before-completion.md` |
| 50 | |
| 51 | ## Quick Decision Tree |
| 52 | |
| 53 | ``` |
| 54 | SITUATION? |
| 55 | │ |
| 56 | ├─ Received feedback |
| 57 | │ ├─ Unclear items? → STOP, ask for clarification first |
| 58 | │ ├─ From human partner? → Understand, then implement |
| 59 | │ └─ From external reviewer? → Verify technically before implementing |
| 60 | │ |
| 61 | ├─ Completed work |
| 62 | │ ├─ Major feature/task? → Request code-reviewer subagent review |
| 63 | │ └─ Before merge? → Request code-reviewer subagent review |
| 64 | │ |
| 65 | └─ About to claim status |
| 66 | ├─ Have fresh verification? → State claim WITH evidence |
| 67 | └─ No fresh verification? → RUN verification command first |
| 68 | ``` |
| 69 | |
| 70 | ## Receiving Feedback Protocol |
| 71 | |
| 72 | ### Response Pattern |
| 73 | READ → UNDERSTAND → VERIFY → EVALUATE → RESPOND → IMPLEMENT |
| 74 | |
| 75 | ### Key Rules |
| 76 | - ❌ No performative agreement: "You're absolutely right!", "Great point!", "Thanks for [anything]" |
| 77 | - ❌ No implementation before verification |
| 78 | - ✅ Restate requirement, ask questions, push back with technical reasoning, or just start working |
| 79 | - ✅ If unclear: STOP and ask for clarification on ALL unclear items first |
| 80 | - ✅ YAGNI check: grep for usage before implementing suggested "proper" features |
| 81 | |
| 82 | ### Source Handling |
| 83 | - **Human partner:** Trusted - implement after understanding, no performative agreement |
| 84 | - **External reviewers:** Verify technically correct, check for breakage, push back if wrong |
| 85 | |
| 86 | **Full protocol:** `references/code-review-reception.md` |
| 87 | |
| 88 | ## Requesting Review Protocol |
| 89 | |
| 90 | ### When to Request |
| 91 | - After each task in subagent-driven development |
| 92 | - After major feature completion |
| 93 | - Before merge to main |
| 94 | |
| 95 | ### Process |
| 96 | 1. Get git SHAs: `BASE_SHA=$(git rev-parse HEAD~1)` and `HEAD_SHA=$(git rev-parse HEAD)` |
| 97 | 2. Dispatch code-reviewer subagent via Task tool with: WHAT_WAS_IMPLEMENTED, PLAN_OR_REQUIREMENTS, BASE_SHA, HEAD_SHA, DESCRIPTION |
| 98 | 3. Act on feedback: Fix Critical immediately, Important before proceeding, note Minor for later |
| 99 | |
| 100 | **Full protocol:** `references/requesting-code-review.md` |
| 101 | |
| 102 | ## Verification Gates Protocol |
| 103 | |
| 104 | ### The Iron Law |
| 105 | **NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE** |
| 106 | |
| 107 | ### Gate Function |
| 108 | IDENTIFY command → RUN full command → READ output → VERIFY confirms claim → THEN claim |
| 109 | |
| 110 | Skip any step = lying, not verifying |
| 111 | |
| 112 | ### Requirements |
| 113 | - Tests pass: Test output shows 0 failures |
| 114 | - Build succeeds: Build command exit 0 |
| 115 | - Bug fixed: Test original symptom passes |
| 116 | - Requirements met: Line-by-line checklist verified |
| 117 | |
| 118 | ### Red Flags - STOP |
| 119 | Using "should"/"probably"/"seems to", expressing satisfaction before verification, committing without verification, trusting agent reports, ANY wording implying success without running verification |
| 120 | |
| 121 | **Full protocol:** `references/verification-before-completion.md` |
| 122 | |
| 123 | ## Integration with Workflows |
| 124 | |
| 125 | - **Subagent-Driven:** Review after EACH task, verify before moving to next |
| 126 | - **Pull Requests:** Verify tests pass, request code-reviewer review before merge |
| 127 | - **General:** Apply verificati |