메뉴
Guides technical evaluation of code review feedback before implementation. Use when receiving PR comments, review suggestions, GitHub feedback, or when asked to address reviewer feedback. Emphasizes verification and reasoned pushback over blind agreement.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Creates TDD implementation plans where tests ARE the plan. Specifies test specs and structural context per phase — implementation emerges during execution, not planning. Verification is fully automated; no manual verification step. Use when planning features test-first, creating TDD plans from design artifacts, or when the user asks for a test-driven implementation plan. NOT for executing a plan test-first (the Red-Green-Refactor cycle during implementation) — that is practicing-tdd.
Breaks an epic into an ordered backlog of vertical, sprint-sized user stories. Use when splitting an epic into stories, planning a feature's story backlog, or when a story feels too large but isn't ready for task breakdown. Produces 4-8 demoable stories, each independently shippable and grounded in user-visible behavior. NOT for sub-story implementation slicing (see slicing-elephant-carpaccio) or for writing one story's full BDD spec (see writing-agile-stories).
Slices a single feature or story into ultra-thin vertical increments (minutes-to-hours each) using Alistair Cockburn's Elephant Carpaccio methodology. Use during implementation planning when one story or feature is too large to build in a session and you need demoable increments across layers. Produces an ordered backlog of 10-20 thin slices, each independently working, testable, and demoable. NOT for splitting an epic into sprint-sized stories (see decomposing-epics).
Write well-scoped, verifiable development tasks for non-user-facing work — refactors, test work, and dependency/tooling changes. Use to "write a dev task", create a "refactor task", a "tech-debt ticket", a "test task", or a "dependency upgrade task"; to capture "non-user-facing work" or a "technical task"; or to write a "definition of done for a refactor". Produces an implementation-focused task with a verifiable Definition of Done. For user-facing behavior, use writing-agile-stories instead.
Expert guidance for creating, writing, and refining Claude Code Skills. Use when working with SKILL.md files, authoring new skills, improving existing skills, or understanding skill structure and best practices.
Examines a git-machete stack of dependent branches and drafts preemptive PR comments warning reviewers that code in an earlier PR is already refactored, renamed, moved, deleted, or fixed later in the stack. Output is DRAFT ONLY — the user pastes it manually; the skill never posts to GitHub. Use for stacked PRs, a git-machete stack, when reviewers comment on code that is "already fixed/refactored/handled later in the stack", to preempt reviewer comments on superseded code, to draft PR comments for a branch stack, or to find code in an earlier branch that a later branch supersedes.
| name | receiving-code-review |
| description | Guides technical evaluation of code review feedback before implementation. Use when receiving PR comments, review suggestions, GitHub feedback, or when asked to address reviewer feedback. Emphasizes verification and reasoned pushback over blind agreement. |
Before implementing any review feedback:
No performative agreement. Just verify, then act.
Code review requires technical evaluation, not emotional performance.
Core principle: Verify before implementing. Ask before assuming. Technical correctness over social comfort.
WHEN receiving code review feedback:
1. READ: Complete feedback without reacting
2. UNDERSTAND: Restate requirement in own words (or ask)
3. VERIFY: Check against codebase reality
4. EVALUATE: Technically sound for THIS codebase?
5. RESPOND: Technical acknowledgment or reasoned pushback
6. IMPLEMENT: One item at a time, test each
NEVER:
INSTEAD:
IF any item is unclear:
STOP - do not implement anything yet
ASK for clarification on unclear items
WHY: Items may be related. Partial understanding = wrong implementation.
Example:
User: "Fix 1-6"
You understand 1,2,3,6. Unclear on 4,5.
❌ WRONG: Implement 1,2,3,6 now, ask about 4,5 later
✅ RIGHT: "I understand items 1,2,3,6. Need clarification on 4 and 5 before proceeding."
BEFORE implementing:
1. Check: Technically correct for THIS codebase?
2. Check: Breaks existing functionality?
3. Check: Reason for current implementation?
4. Check: Works on all platforms/versions?
5. Check: Does reviewer understand full context?
IF suggestion seems wrong:
Push back with technical reasoning
IF can't easily verify:
Say so: "I can't verify this without [X]. Should I [investigate/ask/proceed]?"
IF conflicts with the user's prior decisions:
Stop and discuss with the user first
Guiding principle: External feedback - be skeptical, but check carefully.
IF reviewer suggests "implementing properly":
grep codebase for actual usage
IF unused: "This endpoint isn't called. Remove it (YAGNI)?"
IF used: Then implement properly
Guiding principle: If the feature isn't needed, don't add it - regardless of who suggested it.
FOR multi-item feedback:
1. Clarify anything unclear FIRST
2. Then implement in this order:
- Blocking issues (breaks, security)
- Simple fixes (typos, imports)
- Complex fixes (refactoring, logic)
3. Test each fix individually
4. Verify no regressions
Push back when:
How to push back:
When multiple reviewers suggest contradictory approaches:
Example:
Reviewer A suggests extracting to a helper function.
Reviewer B prefers keeping it inline for readability.
"Two conflicting suggestions: extract vs inline. Extraction adds reusability
but increases indirection. Inline keeps it readable but duplicates logic
if used elsewhere. Which approach do you prefer?"
Never: Implement one reviewer's suggestion while ignoring the other without discussing.
When feedback IS correct:
✅ "Fixed. [Brief description of what changed]"
✅ "Good catch - [specific issue]. Fixed in [location]."
✅ [Just fix it and show in the code]
❌ "You're absolutely right!"
❌ "Great point!"
❌ "Thanks for catching that!"
❌ Excessive gratitude or validation
Why avoid gratitude: Actions speak louder. The fix itself demonstrates you understood the feedback. Focus on what changed, not on thanking.
If you pushed back and were wrong:
✅ "You were right - I checked [X] and it does [Y]. Implementing now."
✅ "Verified this and you're correct. My initial understanding was wrong because [reason]. Fixing."
❌ Long apology
❌ Defending why you pushed back
❌ Over-explaining
State the correction factually and move on.
| Mistake | Fix |
|---|---|
| Performative agreement | State requirement or just act |
| Blind implementation | Verify against codebase first |
| Batch without testing | One at a time, test each |
| Assuming reviewer is right | Check if breaks things |
| Avoiding pushback | Technical correctness > comfort |
| Partial implementation | Clarify all items first |
| Can't verify, proceed anyway | State limitation, ask for direction |
Performative Agreement (Bad):
Reviewer: "Remove legacy code"
❌ "You're absolutely right! Let me remove that..."
Technical Verification (Good):
Reviewer: "Remove legacy code"
✅ "Checking... build target is 10.15+, this API needs 13+. Need legacy for backward compat. Current impl has wrong bundle ID - fix it or drop pre-13 support?"
YAGNI (Good):
Reviewer: "Implement proper metrics tracking with database, date filters, CSV export"
✅ "Grepped codebase - nothing calls this endpoint. Remove it (YAGNI)? Or is there usage I'm missing?"
Unclear Item (Good):
User: "Fix items 1-6"
You understand 1,2,3,6. Unclear on 4,5.
✅ "Understand 1,2,3,6. Need clarification on 4 and 5 before implementing."
When replying to inline review comments on GitHub, reply in the comment thread (gh api repos/{owner}/{repo}/pulls/{pr}/comments/{id}/replies), not as a top-level PR comment.
External feedback = suggestions to evaluate, not orders to follow.
Verify. Question. Then implement.
No performative agreement. Technical rigor always.