Ensure accessibility in UI components including semantic HTML, ARIA attributes, keyboard navigation, and WCAG 2.2 AA compliance.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Two-agent cleanup pattern. Implementer works freely, then a dedicated cleanup agent removes unused imports, console.logs, commented-out code, dead code, and formatting issues. Produces a diff of what was cleaned.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
설치 방법 보기설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.
De-Sloppify implements a two-agent pattern for code cleanup:
This separation prevents "cleanup anxiety" from slowing down implementation while still producing clean final code.
Skill({ skill: 'de-sloppify' });
Invoke when:
CLEANUP AGENT MUST NOT CHANGE BEHAVIOR.
Every removal must be verifiable as dead code, unused import,
or formatting-only. If there is any doubt — LEAVE IT.
| Category | Examples | Safe to Remove |
|---|---|---|
| Unused imports | import foo from 'foo' — never referenced | YES |
| console.log / debug | console.log('debug here'), console.error | YES |
| Commented-out code | // const old = foo() (code, not docs) | YES |
| Dead code | Unreachable branches, unused variables | YES |
| Trailing whitespace | Spaces at end of lines | YES |
| Extra blank lines | 3+ consecutive blank lines | YES |
| TODO with no context | // TODO (no ticket, no description) | On flag only |
| Category | Why |
|---|---|
| Commented doc blocks | /** ... */ JSDoc, /* ... */ section notes |
console.error in catch | May be intentional error logging |
| Disabled test blocks | // it.skip(...) — intentional |
| Feature-flag dead code | May be awaiting activation |
| Complex logic | Even if it looks unused — leave it |
The implementer completes the feature without cleanup overhead. This is normal development.
After implementation, invoke de-sloppify. The cleanup agent runs these checks in order:
Command:
git diff HEAD -- {{target_files}} > C:/dev/projects/agent-studio/.claude/context/tmp/pre-cleanup-snapshot.diff
Expected output: A diff file capturing current state.
Verify: File exists at .claude/context/tmp/pre-cleanup-snapshot.diff.
Command (JavaScript/TypeScript):
node C:/dev/projects/agent-studio/.claude/skills/de-sloppify/scripts/main.cjs \
--action find-unused-imports \
--files "{{comma_separated_file_paths}}"
Expected output: JSON array of { file, line, import } objects for unused imports.
Verify: Exit code 0 and valid JSON array.
Command:
node C:/dev/projects/agent-studio/.claude/skills/de-sloppify/scripts/main.cjs \
--action find-console-logs \
--files "{{comma_separated_file_paths}}"
Expected output: JSON array of { file, line, statement } objects.
Verify: Exit code 0 and valid JSON array.
Command:
node C:/dev/projects/agent-studio/.claude/skills/de-sloppify/scripts/main.cjs \
--action find-commented-code \
--files "{{comma_separated_file_paths}}"
Expected output: JSON array of candidate commented-out code blocks.
Verify: Exit code 0 and valid JSON array.
For each finding that is safe to remove, use Edit to remove the line(s). Apply changes conservatively — when in doubt, leave it.
Command:
cd C:/dev/projects/agent-studio && pnpm lint:fix && pnpm format
Expected output: 0 errors, 0 warnings from lint; no changes from format.
Verify: Both commands exit 0 with no output indicating issues.
Command:
git diff -- {{target_files}}
Expected output: A diff showing only removals (lines starting with -). No additions should appear unless they are indentation fixes.
Verify: Review the diff — confirm every removed line was dead code or formatting. If any functional line was removed, revert with git checkout -- <file>.
Produce a cleanup summary:
De-Sloppify Report
==================
Files processed: N
Unused imports removed: N
Console.logs removed: N
Commented-out code blocks removed: N
Dead code blocks removed: N
Formatting fixes applied: N
Diff: git diff HEAD -- <files>
console.error inside a catch block without confirming it is debug-onlyInput validated against schemas/input.schema.json before execution.
Output contract defined in schemas/output.schema.json.
Pre-execution hook (hooks/pre-execute.cjs) validates:
action is a known enum valuefiles is a non-empty array of pathsPost-execution hook (hooks/post-execute.cjs) emits observability event to .claude/context/runtime/tool-events.jsonl.
developer — runs cleanup after implementationcode-simplifier — incorporates de-sloppify as a cleanup stepcode-reviewer — recommends de-sloppify when reviewing slop-heavy PRsqa — verifies cleanup did not alter behaviorBefore starting: Read .claude/context/memory/learnings.md for cleanup patterns specific to this codebase.
After completing: Append summary to .claude/context/memory/learnings.md:
**de-sloppify** — [date] Cleaned N files. Removed: M unused imports, K console.logs, J commented blocks.
ASSUME INTERRUPTION: Your context may reset. If it's not in memory, it didn't happen.
code-simplifier — structural refactoring (de-sloppify is style-only)finishing-a-development-branch — includes de-sloppify as a phasecodebase-cleaner — broader cleanup including file-level slop아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.
로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.
아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.
Ensure accessibility in UI components including semantic HTML, ARIA attributes, keyboard navigation, and WCAG 2.2 AA compliance.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Use when you want to improve response quality through meta-cognitive reasoning. Applies 15+ reasoning methods to reconsider and refine initial outputs.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
N-round opposing-stance debates for trade-off analysis. Assigns pro/con roles to agents, runs structured debate rounds with quality scoring, and produces a moderator synthesis with confidence-rated recommendation. Generalizable to architecture, technology, security, and design decisions.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Force adversarial code review stance that eliminates confirmation bias — reviewer must find issues or re-analyze
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Creates specialized AI agents on-demand when no existing agent matches a request. Use when the Router cannot find a suitable agent for a task. Enables self-evolution by generating persistent agents.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
LLM-as-judge evaluation framework with 5-dimension rubric (accuracy, groundedness, coherence, completeness, helpfulness) for scoring AI-generated content quality with weighted composite scores and evidence citations
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.