메뉴
Workflow for investigating and reproducing bugs before fixing via TDD. Use when a bug report or GitHub issue is received. Covers issue fetching, root cause analysis, reproduction planning, and handoff to TDD cycle.
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 | code-bug-investigation |
| description | Workflow for investigating and reproducing bugs before fixing via TDD. Use when a bug report or GitHub issue is received. Covers issue fetching, root cause analysis, reproduction planning, and handoff to TDD cycle. |
Structured workflow for going from bug report to reproduction-ready TDD plan.
This is the skill. Everything else is mechanical. Before investigating anything, establish a tight loop that lets you observe the bug:
If you can't reproduce → stop and report to the user immediately. No guessing.
From GitHub issue:
gh issue view {number} --json title,body,labels,assignees,comments
Extract: title, description, steps to reproduce, expected vs actual behavior.
From user description:
Summarize: what's happening, what should happen, reproduction conditions.
Before writing any code, present findings:
## Bug Analysis
**Issue:** {title or summary}
**Root cause:** {explanation}
**Affected code:** {file}:{line range}: {what's wrong}
**Existing coverage:** {tests that exist but missed the bug}
**Reproduction plan:**
| # | Test behavior | Level | Justification |
|---|---------------|-------|---------------|
| 1 | Should_{X}_When_{Y} | Unit | {why} |
Proceed with reproduction test? (Y/n)
Wait for user confirmation before proceeding to TDD cycle.
After analysis is confirmed:
pnpm test, dotnet test, reproduce failuresgh CLI: Fetch issue details, comments, labelsAfter fix is complete:
## Bug Fix Complete
**Issue:** {title}
**Root cause:** {one-line explanation}
**Red:** {test_file}::{test_name}: reproduced ✓
**Green:** {production_file}: {what was fixed}
**Refactor:** {what improved, or "no changes needed"}
**Tests:** pnpm test: {N} passed, 0 failed
CRAP score quality gate for code complexity and test coverage. Use when reviewing methods for complexity risk, deciding whether to split methods, or determining minimum test coverage requirements.
Code smell detection and safe incremental refactoring patterns. Use when improving code structure, reducing technical debt, or cleaning up after a TDD cycle. Covers SOLID principles, smell catalog, and the propose-apply workflow.
Five independent analysis perspectives for code review: correctness, architecture, security, conventions, and complexity. Use when reviewing PRs, commits, or local changes. Includes severity model and output format.
Index of Architecture Decision Records (ADRs) for cross-cutting technical decisions. Use when making design choices, reviewing proposals for consistency with prior decisions, or onboarding to understand why specific tools or patterns were chosen. Covers test tooling, SDK architecture, code quality, and acceptance test infrastructure.
Git commit messages, PR workflow, and branching strategy using Conventional Commits and Semantic Versioning. Use when creating commits, pull requests, or managing Git workflow.
Security guardrails for Envilder (CLI, GitHub Action, SDKs, CDK, website). Covers secret handling, credential hygiene, supply chain safety, input validation, and CI/CD security. Use when reviewing code for security, handling secrets, validating CLI input, or reviewing GitHub Actions workflows.