Doyajin174

@doyajin174 source repo

14 skills
0 followers
1 week ago last updated

▌
Code Migration · doyajin174
Orchestrate system A → system B replacement using scan results, leaf-first dependency order, Build-Check Loop, and module-batch git checkpoints. Handles partial failures gracefully. Use when swapping billing systems, replacing libraries, or removing entire feature modules. MUST trigger on: "마이그레이션", "교체", "갈아끼우기", "시스템 교체", "swap", "migration", "replace system", "remove system", "시스템 제거", "완전 교체".

1 install
▌
Spec Generator · doyajin174
Translate research decisions into concrete implementation specifications. Bridges the gap between "/result (what to use)" and "/guide (build it)". Outputs: target solution, file change map, interfaces, constraints, acceptance criteria. Pipeline position: /question → /result → /research → /result → /spec → /guide. MUST trigger on: "명세", "스펙", "spec", "뭘 만들지 확정", "구현 계획", "implementation spec", "요구사항 정리", "만들 거 정리", "acceptance criteria", or when research is done and the user wants to define exactly what to build before coding starts. NOT for broad exploration (/question), deep research (/research), or actual coding (/guide). State routing: research complete, decisions ready to lock before implementation. If too many unknowns → /research. If impl ready → /guide (after spec). If bug found → /problem. If artifact needs quality review → /validation.

1 install
▌
DB Safety Setup · doyajin174
Set up safe DB migration pipeline for Prisma projects. Adds shadow DB, CI migration checks, deploy pipeline, and documentation automatically. Use when: "DB 세팅", "마이그레이션 세팅", "DB 안전", "db safety", "migration setup", "DB 파이프라인", "prisma migrate 전환".

1 install
▌
AI Only Debugging · doyajin174
Use when debugging any bug — rendering, data pipeline, layout, API, state, async. AI leads with static code analysis, type tracing, git history, and test reproduction. Runtime observation is a structured escalation, not the starting point. Includes layout debugging protocol with arithmetic verification.

1 install
▌
Result Synthesizer · doyajin174
Synthesize findings from multiple sources (internal investigation + external AI responses) into a unified research report. Final step in the prompt pipeline: question (broad) → research (deep) → result (synthesize). MUST trigger on: "결과 종합", "result", "종합해줘", "합쳐줘", "synthesize", "크로스체크", "교차 검증", "결과 정리", "외부 결과", "리서치 결과 정리", or when the user has gathered responses from external AI and wants them combined with internal findings into a final report. State routing: external AI responses received, ready to combine with internal findings. Reusable after any /question or /research round. Routes to: /spec (decisions ready to lock), /guide (trivial tasks), /research (needs more depth), /problem (bug found). Max 2 synthesis rounds per topic before forcing /spec commit.

1 install
▌
Exhaustive Code Scanner · doyajin174
Find ALL references to a code system (enums, types, functions, modules) in a TypeScript/JS codebase with near-100% completeness using a 4-layer pipeline (ripgrep → ast-grep → tsc → knip). Use when removing a system, auditing usage, checking impact, or preparing for migration. MUST trigger on: "스캔", "레퍼런스 찾아", "어디서 쓰이나", "임팩트 분석", "dead code", "다 찾아", "사용처", "참조", "scan", "find all references", "impact analysis", "usage audit".

1 install
▌
Problem Prompt Generator · doyajin174
Generate investigation-to-fix prompts for expected-vs-actual mismatches: runtime errors, build failures, performance regressions. Creates external AI prompts AND runs internal investigation in parallel. Use when behavior differs from expectation — not for design flaws (use /research) or quality review (use /validation). MUST trigger on: "버그", "에러", "안돼", "깨졌어", "크래시", "crash", "bug", "error", "왜 안돼", "fix", "고장", "느려", "slow", "performance issue", "expected X but got Y", "작동 안 함", "실패", "fail", or any request where something was working and now isn't, or where output differs from expected result. NOT for: design flaws or architectural concerns (→ /research), quality review (→ /validation), tech exploration (→ /question). State routing: expected-vs-actual MISMATCH exists (something broken/slow/wrong). If design flaw → /research. If quality review → /validation. If exploring → /question.

1 install
▌
Question Prompt Generator · doyajin174
Explore tech landscape: generates prompts for external AI AND runs internal investigation in parallel. Use BEFORE implementation when the user knows WHAT to build but not HOW. Discovers technologies, trends, and option comparisons. First step in pipeline: question (broad) → research (deep) → result (synthesize). MUST trigger on: "뭐가 있어", "옵션 뭐가 있지", "어떻게 만들지", "방법이 뭐야", "기술 스택", "트렌드", "question", "탐색", "뭘 써야해", "세상에 뭐가 있어", "선택지", "landscape", "what options", "what's available", "tech stack options", or any request where the user has a goal but hasn't decided on an approach yet. NOT for deep-dive on a chosen tech (use research-prompt-generator). State routing: DEFAULT entry when no approach exists. If approach already chosen → /research. If artifact exists → /validation. If bug/error → /problem.

1 install
▌
Research Prompt Generator · doyajin174
Generate structured research prompts for external AI tools (ChatGPT, Gemini, Perplexity, etc.). Auto-reads project context, classifies research type, applies type-specific template, and outputs a self-contained prompt with evidence pack. MUST trigger on: "리서치 프롬프트", "research prompt", "외부 AI", "기술 조사", "비교 분석", "타당성 조사", "feasibility", "technology comparison", "조사해줘", "리서치", or any request to generate a prompt for pasting into an external AI. NOT for bug investigation or fix prompts (use problem-prompt-generator). State routing: user has 1-2 candidate approaches needing deep-dive. If no approach yet → /question. If artifact exists → /validation. If bug/error → /problem.

1 install
▌
Validation Prompt Generator · doyajin174
Generate implementation review prompts for external AI AND run internal code audit in parallel. Use AFTER implementation to evaluate quality, find gaps, and get improvement suggestions. Last step in the development pipeline: question → research → result → guide → validation. MUST trigger on: "평가해줘", "리뷰해줘", "validation", "검증", "이거 괜찮아?", "잘 만든 거 맞아?", "개선점", "어떻게 개선", "수준이 어때", "프로덕션 수준", "production ready", "review my implementation", "code review", "이거 어때", "잘 했어?", "놓친 거 없어?", or any request to evaluate completed implementation quality. NOT for bug fixing (use problem-prompt-generator) or pre-implementation research (use question/research-prompt-generator). State routing: implementation artifact EXISTS and needs quality evaluation. If no artifact yet → /question or /guide first. If bug/mismatch → /problem.

1 install
▌
Guide · doyajin174
Pipeline bootstrap compiler. Reads claude_guide/ reference documents and project CLAUDE.md, then generates skill-specific enriched prompt files for downstream skills. Called once per pipeline by orchestrator before any other skill. Outputs enriched prompt files to .claude/enriched/. State routing: orchestrator bootstrap gate triggers this automatically for multi-skill pipelines. If implementation needed → /implementer. If bug found → /problem.

1 install
▌
Implementer · doyajin174
Execute code implementation using enriched prompts from guide. Dispatches subagents, runs Build-Check Loops, manages git checkpoints. The executor that guide used to be. Use when actual code writing/modification is needed after planning is complete. MUST trigger on: orchestrator delegation for implementation, "구현", "코드 작성", "implement".

1 install
▌
Orchestrator · doyajin174 bundle
Single entry point for all tasks. Classifies the user's request, selects the right skill, invokes it via Skill tool, tracks progress in a state file, and reports back. User-invocable only — invoke via /orchestrator command. Do NOT auto-trigger on generic words like "해줘" or "만들어". This skill is a PURE ROUTER — it NEVER writes code, fixes bugs, or does research itself.

1 install
▌
Design System · doyajin174
Build UI with a systematic design system pipeline — tokens, primitives, components, patterns, pages. Use when building any web interface, dashboard, or app that needs consistent, maintainable design. Enforces "no magic numbers, no eye-balling, no copy-paste" discipline. Triggers on UI/frontend work, component libraries, design tokens, style systems, or when user says "디자인 시스템", "체계적으로 만들어".

1 install

14 published skills