| allowed-tools | ["Read","Write","Grep","Glob","Bash"] |
| description | Use when debugging bugs to find root causes, or after 3 failed fix attempts to stop and document state. |
| name | debugger |
| trigger | 버그 디버깅, 오류 원인 찾기, 에러 추적, root cause, unexpected behavior, bug investigation, 3 strike rule, 가설 검증, hypothesis testing, meta debugging, self code review, debug memory search, 과거 디버그 기록, systematic investigation, 변수 변경 테스트, 편향 피하기, 3 회 실패 후 중단, fresh session, 상태 기록, root cause 분석, 증빙 기반 디버깅 |
Quick Reference
- 3-Strike Rule: 3회 실패 시 즉시 중단 → STATE.md 기록 → fresh session 권장
- Memory First: 조사 전
.hxsk/memories/ 검색으로 과거 유사 사례 확인
- Valid Hypothesis: 반증 가능해야 함 (구체적 원인 명시, 추측 금지)
- One Variable: 한 번에 하나의 변수만 변경하여 테스트 및 문서화
- Persist Findings: 원인 발견 시
root-cause, 배제 시 debug-eliminated 저장
Core Philosophy
User = Reporter, AI = Investigator
User knows: what they expected, what happened, error messages, when it started.
User does NOT know (don't ask): what's causing it, which file, what the fix is.
Ask about experience. Investigate the cause yourself.
Meta-Debugging: Your Own Code
When debugging code you wrote, you're fighting your own mental model.
- Treat your code as foreign — Read it as if someone else wrote it
- Question your design decisions — Your implementations are hypotheses
- Admit your mental model might be wrong — Code behavior is truth
- Prioritize code you touched — Modified lines are prime suspects
Foundation Principles
- What do you know for certain? Observable facts, not assumptions
- What are you assuming? "This library should work this way" — verified?
- Strip away everything you think you know. Build understanding from facts.
Cognitive Biases to Avoid
| Bias | Trap | Antidote |
|---|
| Confirmation | Only look for supporting evidence | Actively seek disconfirming evidence |
| Anchoring | First explanation becomes anchor | Generate 3+ hypotheses before investigating |
| Availability | Recent bugs → assume similar cause | Treat each bug as novel |
| Sunk Cost | Spent 2 hours, keep going | Every 30 min: "Would I still take this path?" |
Systematic Investigation
Change one variable: Make one change, test, observe, document, repeat.
Complete reading: Read entire functions, not just "relevant" lines.
Embrace not knowing: "I don't know" = good (now you can investigate). "It must be X" = dangerous.
Debugging Techniques + Hypothesis Testing + When to Restart → references/debugging-techniques.md
3-Strike Rule
After 3 failed fix attempts:
- STOP the current approach
- Document what was tried in DEBUG.md
- Summarize to STATE.md
- Recommend fresh session with new context
A fresh context often immediately sees what polluted context cannot.
Debug Memory
Before any investigation, search past context:
Grep(pattern: "{symptom}", path: ".hxsk/memories/", output_mode: "files_with_matches")
Persist findings after each session (elimination, root cause, blocked state).
Debug Memory 전체 프로토콜 → references/debug-memory.md
관련 스킬
- REQUIRED:
empirical-validation — 수정 후 경험적 증거로 검증
- REQUIRED:
memory-protocol — 근본 원인, 배제된 가설을 메모리에 저장
- RECOMMENDED:
impact-analysis — 수정 전 영향 범위 분석
Scripts
(없음 — Bash, Read, Grep 등 에이전트 네이티브 도구로 직접 수행)
Iron Laws
NO CONTINUE_AFTER_3_FAILURES WITHOUT STATE_RECORDING_AND_FRESH_SESSION FIRST
NO HYPOTHESIS_FORMULATION WITHOUT FALSIFIABILITY FIRST
NO CHANGE_WITHOUT_ISOLATING_SINGLE_VARIABLE FIRST
NO DIAGNOSIS_WITHOUT_COMPLETE_FUNCTION_READING FIRST
NO INVESTIGATION_WITHOUT_PAST_MEMORY_SEARCH FIRST
NO ASKING_USER_FOR_ROOT_CAUSE FIRST
NO INVESTIGATION_WITHOUT_3_PLUS_HYPOTHESES FIRST
NO DEBUGGING_WITHOUT_TREATING_CODE_AS_FOREIGN FIRST
NO SESSION_END_WITHOUT_MEMORY_PERSISTENCE FIRST
