Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/deathrashed/agents --skill fix명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
SOC 직업 분류 기준
SKILL.md 표시 중
| name | fix |
| description | Smart fix for bugs, types, tests, UI, CI |
{{#if args}}
Input: {{args}}
Check the first word of "{{args}}":
Quick fix mode - minimal analysis, direct action:
Step 1: Quick Analysis
Step 2: Direct Fix
--fix flagStep 3: Testing
Step 4: Summary Report
✓ Files fixed: 3
✓ Tests: Passing
✓ Lint: Clean
Confidence: High
Fix TypeScript/Dart type errors with full workflow:
Step 1: Type Check Execution
# TypeScript
tsc --noEmit
npm run typecheck
# Dart/Flutter
dart analyze
flutter analyze
Step 2: Error Analysis
Step 3: Fix Implementation
Step 4: Verification Loop
Common Fix Patterns:
| Pattern | Before | After |
|---|---|---|
| Null Safety | user.profile.name | user.profile?.name ?? 'Unknown' |
| Type Annotation | function(items) | function(items: Item[]): number |
| Generic Constraint | <T> | <T extends {id: string}> |
| Union Type | direct access | typeof value === 'string' ? ... |
Quick Fixes:
// Null safety
user.profile?.name ?? 'Unknown'
// Type annotation
function calc(items: CartItem[]): number { }
// Generic constraint
function findById<T extends {id: string}>(items: T[], id: string): T | undefined
// Type guard
if (typeof value === 'string') { value.toUpperCase() }
Complex fix mode for difficult bugs (10-15 min):
Stage 1: Discovery (2-3 min)
Files scanned: 1,247
Files identified: 15
Dependencies mapped: 43
Stage 2: Root Cause Analysis (1-2 min)
Root cause identified: [description]
Stage 3: Research (1-2 min)
Stage 4: Planning (1 min)
Strategy created: plans/fix-[issue].md
Files to modify: 8
Tests to add: 15
Stage 5: Implementation (3-5 min)
Stage 6: Validation (1-2 min)
✓ Unit tests: 45/45
✓ Integration tests: 23/23
✓ Regression tests: 67/67
Coverage: 96%
Use for:
Fix failing tests with full pipeline:
Phase 1: Compile
tsc --noEmit for TypeScriptPhase 2: Run Tests
Phase 3: Debug
Phase 4: Plan
Phase 5: Implement
Phase 6: Re-test
Phase 7: Review
Output:
✗ 3 tests failed
- auth/login.test.js: Expected 200, got 401
Debugging...
Root cause: JWT secret not loaded in test env
Planning fix...
Implementing...
✓ All tests passing (48/48)
Fix UI/styling issues with visual analysis:
Step 1: Visual Analysis
Step 2: Code Location
Step 3: Root Cause Identification
Step 4: Fix Implementation
Step 5: Visual Validation
Common UI Issues:
| Type | Examples |
|---|---|
| Layout | Misalignment, overflow, spacing |
| Responsive | Mobile breaks, tablet overlap |
| Visual | Z-index, colors, animations |
| Interaction | Click/touch, hover, focus |
Quick Fixes:
/* Flexbox alignment */
display: flex;
align-items: center;
justify-content: center;
/* Responsive width */
width: 100%;
max-width: 1200px;
/* Z-index stacking */
position: relative;
z-index: 10;
/* Text overflow */
overflow: hidden;
text-overflow: ellipsis;
white-space: nowrap;
Fix CI/CD failures with full analysis:
Step 1: Log Retrieval
gh run view --log-failed to fetch logsgh run view <run-id> --log-failed
Step 2: Error Analysis
Step 3: Solution Research
Step 4: Implementation
Step 5: Verification
actionlintCommon CI Failures:
| Issue | Solution |
|---|---|
| Dependency lock | npm ci instead of npm install |
| Node version | Specify in .nvmrc |
| Test timeout | Increase timeout limit |
| Missing secrets | Check repo secrets |
Analyze log files with full diagnostic workflow:
Step 1: Log Collection
Step 2: Root Cause Analysis
Step 3: Solution Development
Step 4: Verification
Supported Log Formats:
# Standard logs
2025-10-29T14:23:45.123Z [ERROR] Database connection failed
# JSON logs
{"timestamp":"...","level":"error","message":"Request failed"}
# Stack traces
Error: Cannot read property 'id' of null
at UserController.getProfile (src/user.ts:89:15)
Log Collection Commands:
docker logs <container> > logs.txt # Docker
kubectl logs <pod> > logs.txt # Kubernetes
gh run view <id> --log > logs.txt # GitHub Actions
Fix multiple independent issues in parallel:
Step 1: Parse Issues Extract from numbered/bulleted list:
Input: "1. Button bug 2. API error 3. Typo"
Parsed: 3 independent issues
Step 2: Validate Independence Check files don't overlap:
Issue 1: src/components/Button.tsx
Issue 2: src/api/endpoints.ts
Issue 3: src/components/Footer.tsx
✓ No overlap - safe to parallelize
Step 3: Fix Each Issue Process all issues (conceptually parallel):
Step 4: Aggregate Results
Results:
✓ Issue 1: Fixed - Button.tsx
✓ Issue 2: Fixed - endpoints.ts
✓ Issue 3: Fixed - Footer.tsx
Total: 3 files changed
Usage:
/fix parallel
1. Button not clickable
2. API returns 500
3. Typo in footer
Limits:
Debug the specific issue:
Now process: "{{args}}" {{else}}
Intelligent fix for various issues.
/fix <mode> [details]
| Mode | Description |
|---|---|
/fix fast | Quick lint + format |
/fix types | Fix TypeScript errors |
/fix test | Fix failing tests |
/fix ui <component> | Fix UI/CSS issues |
/fix ci | Fix CI/CD config |
/fix logs [file] | Analyze error logs |
/fix <issue> | Debug specific issue |
/fix fast
/fix types
/fix test
/fix ui Button component not responsive
/fix ci GitHub Actions failing
/fix logs npm-debug.log
/fix The login form is not submitting
Choose a mode! {{/if}}