원클릭으로
fix
Debugs and fixes issues systematically. Use when something is broken, throwing errors, or behaving unexpectedly.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Debugs and fixes issues systematically. Use when something is broken, throwing errors, or behaving unexpectedly.
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
| name | fix |
| description | Debugs and fixes issues systematically. Use when something is broken, throwing errors, or behaving unexpectedly. |
| user-invocable | true |
| triggers | ["fix","debug","broken","error"] |
| allowed-tools | Bash, Read, Write, Edit, Grep, Glob |
| model | opus |
| argument-hint | [error or file] |
question: "What's the issue?"
options:
- { label: "Build error", description: "npm run build fails" }
- { label: "Runtime error", description: "App crashes or throws" }
- { label: "UI bug", description: "Something looks wrong" }
- { label: "I'll describe it", description: "Custom issue" }
Build error:
npm run build 2>&1
# Capture and parse error output
Runtime error:
# Check browser console (if agent-browser available and dev server running)
agent-browser open http://localhost:3000
agent-browser errors
# Or check server logs
npm run dev 2>&1 | tail -50
UI bug:
# Take screenshot (if agent-browser available)
agent-browser open http://localhost:3000/[page]
agent-browser screenshot .claude/screenshots/bug-$(date +%s).png
agent-browser snapshot -i
Note: agent-browser daemon fails on Windows. Use npx playwright open <url> as fallback for visual checks.
Common patterns:
| Error Type | Likely Cause |
|---|---|
Cannot find module | Missing import or package |
Type 'X' is not assignable | TypeScript mismatch |
undefined is not a function | Null/undefined access |
Hydration mismatch | Server/client render diff |
CORS error | API endpoint config |
401 Unauthorized | Auth token issue |
For library-specific errors (unfamiliar API, deprecated method, version-specific behavior), query Context7 before guessing or WebSearching:
mcp__plugin_context7_context7__resolve-library-id({ libraryName: "<lib>", query: "<error description>" })
mcp__plugin_context7_context7__query-docs({ libraryId: "<id>", query: "<specific error or API>" })
Version-pinned docs beat cached training data for libraries that move fast (Next.js, Supabase, Stripe, Trigger.dev).
npm run build
npm run test -- --watchAll=false --passWithNoTests 2>/dev/null | tail -5
# If passes, test the specific feature
If build or feature test fails, return to Step 4 with updated error info. Maximum 3 fix attempts — if still failing after 3 tries, escalate to the user with a clear report of what was tried and what's still broken. Do not silently skip or defer.
After fixing, add a test that would have caught the bug:
# If test file exists for the module, add a test case
# If no test file exists, create one for critical paths (auth, billing, data mutations)
For non-critical code, a manual verification is sufficient. For auth/billing/RLS, a test is mandatory.
Append to .claude/decisions.md:
"## [DATE]: Fixed [issue]
- Root cause: [explanation]
- Solution: [what was changed]
- Files: [list of modified files]"
| Say | Action |
|---|---|
fix build | Run build, auto-fix errors |
fix types | Fix TypeScript errors only |
fix [file] | Focus on specific file |
Missing import:
// Add at top of file
import { Thing } from './path'
Null safety:
// Before
obj.property
// After
obj?.property
Type assertion:
// When type is known but TS doesn't infer
(value as ExpectedType)
Stop and ask user if:
Show token / tool usage stats from the local telemetry log. Use when you want to know "which tools am I burning context on", "which skills are expensive", or "was yesterday's session mostly Read/Grep or actually productive".
Parallel quality audit with 7 specialized agents (Opus). Finds bugs, violations, and quality issues. Use audit for fixes, brainstorm for features.
Manage environment variables with Doppler — auto-install CLI, login, link projects, wrap commands with `doppler run`. Replaces scattered .env files with a hub/spoke architecture.
Scaffolds new projects or onboards existing ones. Detects stack, creates monorepo/single-app, configures strict tooling. Use for greenfield or first-time setup.
Archives completed stories from prd.json to reduce token usage.
Autonomous task execution with testing and security. Works through all tasks without stopping.