[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M72X5JSVW6XAF6W7Y0773KD0":3,"skill-same:01M72X5JSVW6XAF6W7Y0773KD0":48},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":25,"commitSha":26,"skillDirectory":27,"forksCount":28,"repositoryHtmlUrl":29,"repositoryHomepageUrl":6,"ownerHtmlUrl":30,"ownerBlogUrl":31,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M72X5JSVW6XAF6W7Y0773KD0","iterative-retrieval",null,"서브에이전트 컨텍스트 문제를 해결하기 위한 점진적 컨텍스트 검색 개선 패턴","SriramGandhiS\u002FSkills","SriramGandhiS","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F211849986?v=4","01M72WQCW12BYX71C6WN584TTA","Sriram S","01M72WQCWTJ2RD0ET1X31AF1RC",1,"MIT","https:\u002F\u002Fgithub.com\u002FSriramGandhiS\u002FSkills\u002Fblob\u002Fmaster\u002Fdocs\u002Fko-KR\u002Fskills\u002Fiterative-retrieval\u002FSKILL.md",[],[],"2026-11-10 22:39:42",false,"---\nname: iterative-retrieval\ndescription: 서브에이전트 컨텍스트 문제를 해결하기 위한 점진적 컨텍스트 검색 개선 패턴\norigin: ECC\n---\n\n# 반복적 검색 패턴\n\n서브에이전트가 작업을 시작하기 전까지 필요한 컨텍스트를 알 수 없는 멀티 에이전트 워크플로우의 \"컨텍스트 문제\"를 해결합니다.\n\n## 활성화 시점\n\n- 사전에 예측할 수 없는 코드베이스 컨텍스트가 필요한 서브에이전트를 생성할 때\n- 컨텍스트가 점진적으로 개선되는 멀티 에이전트 워크플로우를 구축할 때\n- 에이전트 작업에서 \"컨텍스트 초과\" 또는 \"컨텍스트 누락\" 실패를 겪을 때\n- 코드 탐색을 위한 RAG 유사 검색 파이프라인을 설계할 때\n- 에이전트 오케스트레이션에서 토큰 사용량을 최적화할 때\n\n## 문제\n\n서브에이전트는 제한된 컨텍스트로 생성됩니다. 다음을 알 수 없습니다:\n- 관련 코드가 포함된 파일\n- 코드베이스에 존재하는 패턴\n- 프로젝트에서 사용하는 용어\n\n표준 접근법의 실패:\n- **모든 것을 전송**: 컨텍스트 제한 초과\n- **아무것도 전송하지 않음**: 에이전트가 중요한 정보를 갖지 못함\n- **필요한 것을 추측**: 종종 잘못됨\n\n## 해결책: 반복적 검색\n\n컨텍스트를 점진적으로 개선하는 4단계 루프:\n\n```\n┌─────────────────────────────────────────────┐\n│                                             │\n│   ┌──────────┐      ┌──────────┐            │\n│   │ DISPATCH │─────│ EVALUATE │            │\n│   └──────────┘      └──────────┘            │\n│        ▲                  │                 │\n│        │                  ▼                 │\n│   ┌──────────┐      ┌──────────┐            │\n│   │   LOOP   │─────│  REFINE  │            │\n│   └──────────┘      └──────────┘            │\n│                                             │\n│        Max 3 cycles, then proceed           │\n└─────────────────────────────────────────────┘\n```\n\n### 1단계: DISPATCH\n\n후보 파일을 수집하기 위한 초기 광범위 쿼리:\n\n```javascript\n\u002F\u002F Start with high-level intent\nconst initialQuery = {\n  patterns: ['src\u002F**\u002F*.ts', 'lib\u002F**\u002F*.ts'],\n  keywords: ['authentication', 'user', 'session'],\n  excludes: ['*.test.ts', '*.spec.ts']\n};\n\n\u002F\u002F Dispatch to retrieval agent\nconst candidates = await retrieveFiles(initialQuery);\n```\n\n### 2단계: EVALUATE\n\n검색된 콘텐츠의 관련성 평가:\n\n```javascript\nfunction evaluateRelevance(files, task) {\n  return files.map(file => ({\n    path: file.path,\n    relevance: scoreRelevance(file.content, task),\n    reason: explainRelevance(file.content, task),\n    missingContext: identifyGaps(file.content, task)\n  }));\n}\n```\n\n점수 기준:\n- **높음 (0.8-1.0)**: 대상 기능을 직접 구현\n- **중간 (0.5-0.7)**: 관련 패턴이나 타입을 포함\n- **낮음 (0.2-0.4)**: 간접적으로 관련\n- **없음 (0-0.2)**: 관련 없음, 제외\n\n### 3단계: REFINE\n\n평가를 기반으로 검색 기준 업데이트:\n\n```javascript\nfunction refineQuery(evaluation, previousQuery) {\n  return {\n    \u002F\u002F Add new patterns discovered in high-relevance files\n    patterns: [...previousQuery.patterns, ...extractPatterns(evaluation)],\n\n    \u002F\u002F Add terminology found in codebase\n    keywords: [...previousQuery.keywords, ...extractKeywords(evaluation)],\n\n    \u002F\u002F Exclude confirmed irrelevant paths\n    excludes: [...previousQuery.excludes, ...evaluation\n      .filter(e => e.relevance \u003C 0.2)\n      .map(e => e.path)\n    ],\n\n    \u002F\u002F Target specific gaps\n    focusAreas: evaluation\n      .flatMap(e => e.missingContext)\n      .filter(unique)\n  };\n}\n```\n\n### 4단계: LOOP\n\n개선된 기준으로 반복 (최대 3회):\n\n```javascript\nasync function iterativeRetrieve(task, maxCycles = 3) {\n  let query = createInitialQuery(task);\n  let bestContext = [];\n\n  for (let cycle = 0; cycle \u003C maxCycles; cycle++) {\n    const candidates = await retrieveFiles(query);\n    const evaluation = evaluateRelevance(candidates, task);\n\n    \u002F\u002F Check if we have sufficient context\n    const highRelevance = evaluation.filter(e => e.relevance >= 0.7);\n    if (highRelevance.length >= 3 && !hasCriticalGaps(evaluation)) {\n      return highRelevance;\n    }\n\n    \u002F\u002F Refine and continue\n    query = refineQuery(evaluation, query);\n    bestContext = mergeContext(bestContext, highRelevance);\n  }\n\n  return bestContext;\n}\n```\n\n## 실용적인 예시\n\n### 예시 1: 버그 수정 컨텍스트\n\n```\nTask: \"Fix the authentication token expiry bug\"\n\nCycle 1:\n  DISPATCH: Search for \"token\", \"auth\", \"expiry\" in src\u002F**\n  EVALUATE: Found auth.ts (0.9), tokens.ts (0.8), user.ts (0.3)\n  REFINE: Add \"refresh\", \"jwt\" keywords; exclude user.ts\n\nCycle 2:\n  DISPATCH: Search refined terms\n  EVALUATE: Found session-manager.ts (0.95), jwt-utils.ts (0.85)\n  REFINE: Sufficient context (2 high-relevance files)\n\nResult: auth.ts, tokens.ts, session-manager.ts, jwt-utils.ts\n```\n\n### 예시 2: 기능 구현\n\n```\nTask: \"Add rate limiting to API endpoints\"\n\nCycle 1:\n  DISPATCH: Search \"rate\", \"limit\", \"api\" in routes\u002F**\n  EVALUATE: No matches - codebase uses \"throttle\" terminology\n  REFINE: Add \"throttle\", \"middleware\" keywords\n\nCycle 2:\n  DISPATCH: Search refined terms\n  EVALUATE: Found throttle.ts (0.9), middleware\u002Findex.ts (0.7)\n  REFINE: Need router patterns\n\nCycle 3:\n  DISPATCH: Search \"router\", \"express\" patterns\n  EVALUATE: Found router-setup.ts (0.8)\n  REFINE: Sufficient context\n\nResult: throttle.ts, middleware\u002Findex.ts, router-setup.ts\n```\n\n## 에이전트와의 통합\n\n에이전트 프롬프트에서 사용:\n\n```markdown\nWhen retrieving context for this task:\n1. Start with broad keyword search\n2. Evaluate each file's relevance (0-1 scale)\n3. Identify what context is still missing\n4. Refine search criteria and repeat (max 3 cycles)\n5. Return files with relevance >= 0.7\n```\n\n## 모범 사례\n\n1. **광범위하게 시작하여 점진적으로 좁히기** - 초기 쿼리를 과도하게 지정하지 않기\n2. **코드베이스 용어 학습** - 첫 번째 사이클에서 주로 네이밍 컨벤션이 드러남\n3. **누락된 것 추적** - 명시적 격차 식별이 개선을 주도\n4. **\"충분히 좋은\" 수준에서 중단** - 관련성 높은 파일 3개가 보통 수준의 파일 10개보다 나음\n5. **자신 있게 제외** - 관련성 낮은 파일은 관련성이 높아지지 않음\n\n## 관련 항목\n\n- [The Longform Guide](https:\u002F\u002Fx.com\u002Faffaanmustafa\u002Fstatus\u002F2014040193557471352) - 서브에이전트 오케스트레이션 섹션\n- `continuous-learning` 스킬 - 시간이 지남에 따라 개선되는 패턴\n- `~\u002F.claude\u002Fagents\u002F`의 에이전트 정의\n","# 반복적 검색 패턴\n\n서브에이전트가 작업을 시작하기 전까지 필요한 컨텍스트를 알 수 없는 멀티 에이전트 워크플로우의 \"컨텍스트 문제\"를 해결합니다.\n\n## 활성화 시점\n\n- 사전에 예측할 수 없는 코드베이스 컨텍스트가 필요한 서브에이전트를 생성할 때\n- 컨텍스트가 점진적으로 개선되는 멀티 에이전트 워크플로우를 구축할 때\n- 에이전트 작업에서 \"컨텍스트 초과\" 또는 \"컨텍스트 누락\" 실패를 겪을 때\n- 코드 탐색을 위한 RAG 유사 검색 파이프라인을 설계할 때\n- 에이전트 오케스트레이션에서 토큰 사용량을 최적화할 때\n\n## 문제\n\n서브에이전트는 제한된 컨텍스트로 생성됩니다. 다음을 알 수 없습니다:\n- 관련 코드가 포함된 파일\n- 코드베이스에 존재하는 패턴\n- 프로젝트에서 사용하는 용어\n\n표준 접근법의 실패:\n- **모든 것을 전송**: 컨텍스트 제한 초과\n- **아무것도 전송하지 않음**: 에이전트가 중요한 정보를 갖지 못함\n- **필요한 것을 추측**: 종종 잘못됨\n\n## 해결책: 반복적 검색\n\n컨텍스트를 점진적으로 개선하는 4단계 루프:\n\n```\n┌─────────────────────────────────────────────┐\n│                                             │\n│   ┌──────────┐      ┌──────────┐            │\n│   │ DISPATCH │─────│ EVALUATE │            │\n│   └──────────┘      └──────────┘            │\n│        ▲                  │                 │\n│        │                  ▼                 │\n│   ┌──────────┐      ┌──────────┐            │\n│   │   LOOP   │─────│  REFINE  │            │\n│   └──────────┘      └──────────┘            │\n│                                             │\n│        Max 3 cycles, then proceed           │\n└─────────────────────────────────────────────┘\n```\n\n### 1단계: DISPATCH\n\n후보 파일을 수집하기 위한 초기 광범위 쿼리:\n\n```javascript\n\u002F\u002F Start with high-level intent\nconst initialQuery = {\n  patterns: ['src\u002F**\u002F*.ts', 'lib\u002F**\u002F*.ts'],\n  keywords: ['authentication', 'user', 'session'],\n  excludes: ['*.test.ts', '*.spec.ts']\n};\n\n\u002F\u002F Dispatch to retrieval agent\nconst candidates = await retrieveFiles(initialQuery);\n```\n\n### 2단계: EVALUATE\n\n검색된 콘텐츠의 관련성 평가:\n\n```javascript\nfunction evaluateRelevance(files, task) {\n  return files.map(file => ({\n    path: file.path,\n    relevance: scoreRelevance(file.content, task),\n    reason: explainRelevance(file.content, task),\n    missingContext: identifyGaps(file.content, task)\n  }));\n}\n```\n\n점수 기준:\n- **높음 (0.8-1.0)**: 대상 기능을 직접 구현\n- **중간 (0.5-0.7)**: 관련 패턴이나 타입을 포함\n- **낮음 (0.2-0.4)**: 간접적으로 관련\n- **없음 (0-0.2)**: 관련 없음, 제외\n\n### 3단계: REFINE\n\n평가를 기반으로 검색 기준 업데이트:\n\n```javascript\nfunction refineQuery(evaluation, previousQuery) {\n  return {\n    \u002F\u002F Add new patterns discovered in high-relevance files\n    patterns: [...previousQuery.patterns, ...extractPatterns(evaluation)],\n\n    \u002F\u002F Add terminology found in codebase\n    keywords: [...previousQuery.keywords, ...extractKeywords(evaluation)],\n\n    \u002F\u002F Exclude confirmed irrelevant paths\n    excludes: [...previousQuery.excludes, ...evaluation\n      .filter(e => e.relevance \u003C 0.2)\n      .map(e => e.path)\n    ],\n\n    \u002F\u002F Target specific gaps\n    focusAreas: evaluation\n      .flatMap(e => e.missingContext)\n      .filter(unique)\n  };\n}\n```\n\n### 4단계: LOOP\n\n개선된 기준으로 반복 (최대 3회):\n\n```javascript\nasync function iterativeRetrieve(task, maxCycles = 3) {\n  let query = createInitialQuery(task);\n  let bestContext = [];\n\n  for (let cycle = 0; cycle \u003C maxCycles; cycle++) {\n    const candidates = await retrieveFiles(query);\n    const evaluation = evaluateRelevance(candidates, task);\n\n    \u002F\u002F Check if we have sufficient context\n    const highRelevance = evaluation.filter(e => e.relevance >= 0.7);\n    if (highRelevance.length >= 3 && !hasCriticalGaps(evaluation)) {\n      return highRelevance;\n    }\n\n    \u002F\u002F Refine and continue\n    query = refineQuery(evaluation, query);\n    bestContext = mergeContext(bestContext, highRelevance);\n  }\n\n  return bestContext;\n}\n```\n\n## 실용적인 예시\n\n### 예시 1: 버그 수정 컨텍스트\n\n```\nTask: \"Fix the authentication token expiry bug\"\n\nCycle 1:\n  DISPATCH: Search for \"token\", \"auth\", \"expiry\" in src\u002F**\n  EVALUATE: Found auth.ts (0.9), tokens.ts (0.8), user.ts (0.3)\n  REFINE: Add \"refresh\", \"jwt\" keywords; exclude user.ts\n\nCycle 2:\n  DISPATCH: Search refined terms\n  EVALUATE: Found session-manager.ts (0.95), jwt-utils.ts (0.85)\n  REFINE: Sufficient context (2 high-relevance files)\n\nResult: auth.ts, tokens.ts, session-manager.ts, jwt-utils.ts\n```\n\n### 예시 2: 기능 구현\n\n```\nTask: \"Add rate limiting to API endpoints\"\n\nCycle 1:\n  DISPATCH: Search \"rate\", \"limit\", \"api\" in routes\u002F**\n  EVALUATE: No matches - codebase uses \"throttle\" terminology\n  REFINE: Add \"throttle\", \"middleware\" keywords\n\nCycle 2:\n  DISPATCH: Search refined terms\n  EVALUATE: Found throttle.ts (0.9), middleware\u002Findex.ts (0.7)\n  REFINE: Need router patterns\n\nCycle 3:\n  DISPATCH: Search \"router\", \"express\" patterns\n  EVALUATE: Found router-setup.ts (0.8)\n  REFINE: Sufficient context\n\nResult: throttle.ts, middleware\u002Findex.ts, router-setup.ts\n```\n\n## 에이전트와의 통합\n\n에이전트 프롬프트에서 사용:\n\n```markdown\nWhen retrieving context for this task:\n1. Start with broad keyword search\n2. Evaluate each file's relevance (0-1 scale)\n3. Identify what context is still missing\n4. Refine search criteria and repeat (max 3 cycles)\n5. Return files with relevance >= 0.7\n```\n\n## 모범 사례\n\n1. **광범위하게 시작하여 점진적으로 좁히기** - 초기 쿼리를 과도하게 지정하지 않기\n2. **코드베이스 용어 학습** - 첫 번째 사이클에서 주로 네이밍 컨벤션이 드러남\n3. **누락된 것 추적** - 명시적 격차 식별이 개선을 주도\n4. **\"충분히 좋은\" 수준에서 중단** - 관련성 높은 파일 3개가 보통 수준의 파일 10개보다 나음\n5. **자신 있게 제외** - 관련성 낮은 파일은 관련성이 높아지지 않음\n\n## 관련 항목\n\n- [The Longform Guide](https:\u002F\u002Fx.com\u002Faffaanmustafa\u002Fstatus\u002F2014040193557471352) - 서브에이전트 오케스트레이션 섹션\n- `continuous-learning` 스킬 - 시간이 지남에 따라 개선되는 패턴\n- `~\u002F.claude\u002Fagents\u002F`의 에이전트 정의\n",{"name":5,"origin":24,"description":7},"ECC","master","d6ca01da0f24a3b0e26e4a8b7b66cbd25d429f5d","docs\u002Fko-KR\u002Fskills\u002Fiterative-retrieval",0,"https:\u002F\u002Fgithub.com\u002FSriramGandhiS\u002FSkills","https:\u002F\u002Fgithub.com\u002FSriramGandhiS","https:\u002F\u002Fsriram.website\u002F","Curious guy who keeps tinkering until something works","2026-05-20 16:21:34",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add SriramGandhiS\u002FSkills --skill iterative-retrieval",{"tool":40,"command":41},"bunx","bunx skills add SriramGandhiS\u002FSkills --skill iterative-retrieval",{"tool":43,"command":44},"pnpm","pnpm dlx skills add SriramGandhiS\u002FSkills --skill iterative-retrieval","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FSriramGandhiS\u002FSkills\u002Fblob\u002Fmaster\u002Fdocs\u002Fko-KR\u002Fskills\u002Fiterative-retrieval\u002FSKILL.md",true,[],[49,57,62,68,75,82],{"skillNo":50,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":56},"01M732609EPJWV13PVTX3GT4TM","webgpu-threejs-tsl","Comprehensive guide for developing WebGPU-enabled Three.js applications using TSL (Three.js Shading Language). Covers WebGPU renderer setup, TSL syntax and node materials, compute shaders, post-processing effects, and WGSL integration. Use this skill when working with Three.js WebGPU, TSL shaders, node materials, or GPU compute in Three.js.","https:\u002F\u002Fgithub.com\u002FSriramGandhiS\u002FSkills\u002Fblob\u002Fmaster\u002Fwebgpu-threejs-tsl\u002Fskills\u002Fwebgpu-threejs-tsl\u002FSKILL.md",[],[],"2026-11-11 15:11:14",{"skillNo":58,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":56},"01M7325XM47AGWB6K6C4VXKE6B","https:\u002F\u002Fgithub.com\u002FSriramGandhiS\u002FSkills\u002Fblob\u002Fmaster\u002Fwebgpu-threejs-tsl\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":56},"01M7325WP6DES69JKC840VMFEP","verification-loop","https:\u002F\u002Fgithub.com\u002FSriramGandhiS\u002FSkills\u002Fblob\u002Fmaster\u002Fweb-assets\u002Fdocs\u002Fzh-TW\u002Fskills\u002Fverification-loop\u002FSKILL.md",[],[],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":56},"01M7325TFQF1DZTQ6J3WGQQ1VR","tdd-workflow","Use this skill when writing new features, fixing bugs, or refactoring code. Enforces test-driven development with 80%+ coverage including unit, integration, and E2E tests.","https:\u002F\u002Fgithub.com\u002FSriramGandhiS\u002FSkills\u002Fblob\u002Fmaster\u002Fweb-assets\u002Fdocs\u002Fzh-TW\u002Fskills\u002Ftdd-workflow\u002FSKILL.md",[],[],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":56},"01M7325RGV9AY9PK92D2V8KKPW","strategic-compact","Suggests manual context compaction at logical intervals to preserve context through task phases rather than arbitrary auto-compaction.","https:\u002F\u002Fgithub.com\u002FSriramGandhiS\u002FSkills\u002Fblob\u002Fmaster\u002Fweb-assets\u002Fdocs\u002Fzh-TW\u002Fskills\u002Fstrategic-compact\u002FSKILL.md",[],[],{"skillNo":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":56},"01M7325MZC28SE86E7NMCNEFS4","security-review","Use this skill when adding authentication, handling user input, working with secrets, creating API endpoints, or implementing payment\u002Fsensitive features. Provides comprehensive security checklist and patterns.","https:\u002F\u002Fgithub.com\u002FSriramGandhiS\u002FSkills\u002Fblob\u002Fmaster\u002Fweb-assets\u002Fdocs\u002Fzh-TW\u002Fskills\u002Fsecurity-review\u002FSKILL.md",[],[]]