[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M79A9ZN9D10FQ6MV30DAZPTP":3,"skill-same:01M79A9ZN9D10FQ6MV30DAZPTP":78},{"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":31,"commitSha":32,"skillDirectory":33,"forksCount":14,"repositoryHtmlUrl":34,"repositoryHomepageUrl":6,"ownerHtmlUrl":35,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":36,"install":37,"resources":49},"01M79A9ZN9D10FQ6MV30DAZPTP","google-gemini-file-search",null,"Google Gemini File Search for managed RAG with 100+ file formats. Use for document Q&A, knowledge bases, or encountering immutability errors, quota issues, polling failures. Supports Gemini 3 Pro\u002FFlash (Gemini 2.5 legacy).","Elogic360\u002Fsoftware-factory","Elogic360","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F156535519?v=4","01M74NH3F0RB13PWB7JZDE2V80","elogic360","01M74NH3FPC65082MDZVMFHDM3",0,"MIT","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fintegrations\u002Fclaude-skills-secondsky\u002Fplugins\u002Fgoogle-gemini-file-search\u002Fskills\u002Fgoogle-gemini-file-search\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: google-gemini-file-search\ndescription: Google Gemini File Search for managed RAG with 100+ file formats. Use for document Q&A, knowledge bases, or encountering immutability errors, quota issues, polling failures. Supports Gemini 3 Pro\u002FFlash (Gemini 2.5 legacy).\nlicense: MIT\nmetadata:\n  version: \"2.0.0\"\n  last_verified: \"2025-11-18\"\n  production_tested: true\n  token_savings: \"~65%\"\n  errors_prevented: 8\n  templates_included: 0\n  references_included: 2\n---\n\n# Google Gemini File Search\n\n**Status**: Production Ready | **Last Verified**: 2025-11-18\n\n---\n\n## What Is File Search?\n\nGoogle Gemini File Search is **fully managed RAG** (Retrieval-Augmented Generation):\n- Upload documents → Automatic chunking + embeddings + vector search + citations\n- **No vector database setup** required\n- **100+ file formats** supported (PDF, Word, Excel, code, Markdown, JSON, etc.)\n- **Built-in grounding** with citation metadata\n- **Cost-effective**: $0.15\u002F1M tokens (one-time indexing), free storage + queries\n\n**Key difference from other RAG:**\n- Cloudflare Vectorize: You manage chunking\u002Fembeddings\n- OpenAI Files API: Tied to Assistants API threads\n- File Search: Fully managed, standalone RAG\n\n---\n\n## Quick Start (5 Minutes)\n\n### 1. Get API Key & Install\n\nGet API key: https:\u002F\u002Faistudio.google.com\u002Fapikey (Free tier: 1 GB storage, 1,500 requests\u002Fday)\n\n```bash\nbun add @google\u002Fgenai\n```\n\n**Version:** 0.21.0+ | **Node.js:** 18+\n\n### 2. Basic Example\n\n```typescript\nimport { GoogleGenerativeAI } from '@google\u002Fgenai';\nimport fs from 'fs';\n\nconst ai = new GoogleGenerativeAI(process.env.GOOGLE_AI_API_KEY);\n\n\u002F\u002F Create store\nconst fileStore = await ai.fileSearchStores.create({\n  config: { displayName: 'my-knowledge-base' }\n});\n\n\u002F\u002F Upload document\nconst operation = await ai.fileSearchStores.uploadToFileSearchStore({\n  name: fileStore.name,\n  file: fs.createReadStream('.\u002Fmanual.pdf'),\n  config: {\n    displayName: 'Installation Manual',\n    chunkingConfig: {\n      whiteSpaceConfig: {\n        maxTokensPerChunk: 500,\n        maxOverlapTokens: 50\n      }\n    }\n  }\n});\n\n\u002F\u002F Poll until done\nwhile (!operation.done) {\n  await new Promise(resolve => setTimeout(resolve, 1000));\n  operation = await ai.operations.get({ name: operation.name });\n}\n\n\u002F\u002F Query documents\nconst model = ai.getGenerativeModel({\n  model: 'gemini-2.5-pro',  \u002F\u002F Only 2.5 Pro\u002FFlash supported\n  tools: [{\n    fileSearchTool: {\n      fileSearchStores: [fileStore.name]\n    }\n  }]\n});\n\nconst result = await model.generateContent('How do I install the product?');\nconsole.log(result.response.text());\n\n\u002F\u002F Get citations\nconst grounding = result.response.candidates[0].groundingMetadata;\nif (grounding) {\n  console.log('Sources:', grounding.groundingChunks);\n}\n```\n\n**Load `references\u002Fsetup-guide.md` for complete walkthrough with batch uploads, error handling, and production checklist.**\n\n---\n\n## Critical Rules\n\n### Always Do\n\n1. **Use delete + re-upload** for updates (documents are immutable)\n2. **Calculate 3x storage** (embeddings + metadata = ~3x file size)\n3. **Configure chunking** (500 tokens for technical docs, 800 for prose)\n4. **Poll operations** until `done: true` (with timeout)\n5. **Use force: true** when deleting stores with documents\n6. **Use Gemini 2.5 models** only (2.5-pro or 2.5-flash)\n7. **Keep metadata under 20 fields** per document\n8. **Estimate indexing costs** ($0.15\u002F1M tokens one-time)\n\n### Never Do\n\n1. **Never try to update** documents (no PATCH API exists)\n2. **Never assume storage = file size** (it's 3x)\n3. **Never skip chunking config** (defaults may not be optimal)\n4. **Never upload without polling** (operation may still be processing)\n5. **Never delete without force** if store has documents\n6. **Never use Gemini 1.5 models** (File Search requires 2.5)\n7. **Never exceed 20 metadata fields** (hard limit)\n8. **Never upload large files without cost estimate**\n\n---\n\n## Top 3 Errors Prevented\n\n### Error 1: Document Immutability\n\n**Problem:** Trying to update existing document\n\n**Solution:** Delete + re-upload pattern\n\n```typescript\n\u002F\u002F Find and delete old version\nconst docs = await ai.fileSearchStores.documents.list({\n  parent: fileStore.name\n});\nconst oldDoc = docs.documents.find(d => d.displayName === 'manual.pdf');\nif (oldDoc) {\n  await ai.fileSearchStores.documents.delete({\n    name: oldDoc.name,\n    force: true\n  });\n}\n\n\u002F\u002F Upload new version\nawait ai.fileSearchStores.uploadToFileSearchStore({\n  name: fileStore.name,\n  file: fs.createReadStream('manual-v2.pdf'),\n  config: { displayName: 'manual.pdf' }\n});\n```\n\n### Error 2: Storage Quota Exceeded\n\n**Problem:** Storage calculation wrong (3x multiplier)\n\n**Solution:** Estimate before upload\n\n```typescript\nconst fileSize = fs.statSync('data.pdf').size;\nconst estimatedStorage = fileSize * 3;  \u002F\u002F Embeddings + metadata\n\nif (estimatedStorage > 1e9) {\n  console.warn('⚠️ May exceed free tier 1 GB limit');\n}\n```\n\n### Error 3: Model Compatibility\n\n**Problem:** Using wrong model version\n\n**Solution:** Use Gemini 2.5 only\n\n```typescript\n\u002F\u002F ✅ CORRECT\nconst model = ai.getGenerativeModel({\n  model: 'gemini-2.5-pro',  \u002F\u002F or gemini-2.5-flash\n  tools: [{ fileSearchTool: { fileSearchStores: [storeName] } }]\n});\n\n\u002F\u002F ❌ WRONG\nconst model = ai.getGenerativeModel({\n  model: 'gemini-1.5-pro',  \u002F\u002F Not supported!\n  tools: [{ fileSearchTool: { fileSearchStores: [storeName] } }]\n});\n```\n\n**Load `references\u002Ferror-catalog.md` for all 8 errors with detailed solutions including chunking, operation polling, metadata limits, and force delete requirements.**\n\n---\n\n## When to Use File Search\n\n### Use File Search When:\n\n- Want fully managed RAG (no vector DB)\n- Cost predictability matters (one-time indexing)\n- Need 100+ file format support\n- Citations are important (built-in grounding)\n- Simple deployment is priority\n- Documents are relatively static\n\n### Use Alternatives When:\n\n**Cloudflare Vectorize** - Global edge performance, custom embeddings, real-time R2 updates\n**OpenAI Files API** - Assistants API, conversational threads, very large collections (10,000+)\n\n---\n\n## Common Patterns\n\n### Pattern 1: Customer Support Knowledge Base\n\n```typescript\n\u002F\u002F Upload support docs with metadata\nawait ai.fileSearchStores.uploadToFileSearchStore({\n  name: fileStore.name,\n  file: fs.createReadStream('troubleshooting.pdf'),\n  config: {\n    displayName: 'Troubleshooting Guide',\n    customMetadata: {\n      doc_type: 'support',\n      category: 'troubleshooting',\n      language: 'en'\n    }\n  }\n});\n```\n\n### Pattern 2: Batch Document Upload\n\n```typescript\nconst files = ['doc1.pdf', 'doc2.md', 'doc3.docx'];\nconst uploadPromises = files.map(file =>\n  ai.fileSearchStores.uploadToFileSearchStore({\n    name: fileStore.name,\n    file: fs.createReadStream(file),\n    config: { displayName: file }\n  })\n);\nconst operations = await Promise.all(uploadPromises);\n\n\u002F\u002F Poll all operations\nfor (const op of operations) {\n  let operation = op;\n  while (!operation.done) {\n    await new Promise(resolve => setTimeout(resolve, 1000));\n    operation = await ai.operations.get({ name: operation.name });\n  }\n  console.log('✅', operation.response.displayName);\n}\n```\n\n### Pattern 3: Document Update Flow\n\n```typescript\n\u002F\u002F 1. List existing documents\nconst docs = await ai.fileSearchStores.documents.list({\n  parent: fileStore.name\n});\n\n\u002F\u002F 2. Delete old version\nconst oldDoc = docs.documents.find(d => d.displayName === 'manual.pdf');\nif (oldDoc) {\n  await ai.fileSearchStores.documents.delete({\n    name: oldDoc.name,\n    force: true\n  });\n}\n\n\u002F\u002F 3. Upload new version\nconst operation = await ai.fileSearchStores.uploadToFileSearchStore({\n  name: fileStore.name,\n  file: fs.createReadStream('manual-v2.pdf'),\n  config: {\n    displayName: 'manual.pdf',\n    customMetadata: {\n      version: '2.0',\n      updated_at: new Date().toISOString()\n    }\n  }\n});\n\n\u002F\u002F 4. Poll until done\nwhile (!operation.done) {\n  await new Promise(resolve => setTimeout(resolve, 1000));\n  operation = await ai.operations.get({ name: operation.name });\n}\n```\n\n**Load `references\u002Fsetup-guide.md` for additional patterns including code documentation search and internal knowledge bases.**\n\n---\n\n## When to Load References\n\n### Load `references\u002Fsetup-guide.md` when:\n- First-time File Search setup\n- Need step-by-step walkthrough with all configuration options\n- Configuring batch upload strategies\n- Production deployment checklist\n- Complete API initialization patterns\n\n### Load `references\u002Ferror-catalog.md` when:\n- Encountering any of 8 common errors\n- Need detailed error solutions with code examples\n- Prevention checklist required\n- Troubleshooting upload\u002Fquery issues\n- Understanding chunking, metadata, or cost calculation problems\n\n---\n\n## Supported File Formats\n\n**100+ formats including:**\n- **Documents**: PDF, Word (.docx), Excel (.xlsx), PowerPoint (.pptx)\n- **Text**: Markdown (.md), Plain text (.txt), JSON, CSV\n- **Code**: Python, JavaScript, TypeScript, Java, C++, Go, Rust, etc.\n\n**Not supported:** Images in PDFs (text extraction only), Audio files, Video files\n\n---\n\n## Pricing\n\n**Indexing (one-time):** $0.15 per 1M tokens\n**Storage:** Free (10 GB - 1 TB depending on tier)\n**Query embeddings:** Free (retrieved context counts as input tokens)\n\n**Example:** 1,000-page document ≈ 500k tokens → Indexing cost: $0.075 → Storage: ~1.5 GB (3x multiplier)\n\n---\n\n## Chunking Guidelines\n\n**Technical docs:** 500 tokens\u002Fchunk, 50 overlap\n**Prose:** 800 tokens\u002Fchunk, 80 overlap\n**Legal:** 300 tokens\u002Fchunk, 30 overlap\n\n```typescript\nchunkingConfig: {\n  whiteSpaceConfig: {\n    maxTokensPerChunk: 500,  \u002F\u002F Smaller = more precise\n    maxOverlapTokens: 50     \u002F\u002F 10% overlap recommended\n  }\n}\n```\n\n---\n\n## Resources\n\n**References** (`references\u002F`):\n- `setup-guide.md` - Complete setup walkthrough (authentication, store creation, file upload, batch patterns, production checklist)\n- `error-catalog.md` - All 8 documented errors with solutions (immutability, storage, chunking, metadata, costs, polling, force delete, model compatibility)\n\n**Official Documentation**:\n- **File Search Overview**: https:\u002F\u002Fai.google.dev\u002Fapi\u002Ffile-search\n- **API Reference**: https:\u002F\u002Fai.google.dev\u002Fapi\u002Ffile-search\u002Fdocuments\n- **Blog Post**: https:\u002F\u002Fblog.google\u002Ftechnology\u002Fdevelopers\u002Ffile-search-gemini-api\u002F\n\n---\n\n**Questions? Issues?**\n\n1. Check `references\u002Fsetup-guide.md` for complete setup\n2. Review `references\u002Ferror-catalog.md` for all 8 errors\n3. Verify model version (must be Gemini 2.5)\n4. Check storage calculation (3x file size)\n","# Google Gemini File Search\n\n**Status**: Production Ready | **Last Verified**: 2025-11-18\n\n---\n\n## What Is File Search?\n\nGoogle Gemini File Search is **fully managed RAG** (Retrieval-Augmented Generation):\n- Upload documents → Automatic chunking + embeddings + vector search + citations\n- **No vector database setup** required\n- **100+ file formats** supported (PDF, Word, Excel, code, Markdown, JSON, etc.)\n- **Built-in grounding** with citation metadata\n- **Cost-effective**: $0.15\u002F1M tokens (one-time indexing), free storage + queries\n\n**Key difference from other RAG:**\n- Cloudflare Vectorize: You manage chunking\u002Fembeddings\n- OpenAI Files API: Tied to Assistants API threads\n- File Search: Fully managed, standalone RAG\n\n---\n\n## Quick Start (5 Minutes)\n\n### 1. Get API Key & Install\n\nGet API key: https:\u002F\u002Faistudio.google.com\u002Fapikey (Free tier: 1 GB storage, 1,500 requests\u002Fday)\n\n```bash\nbun add @google\u002Fgenai\n```\n\n**Version:** 0.21.0+ | **Node.js:** 18+\n\n### 2. Basic Example\n\n```typescript\nimport { GoogleGenerativeAI } from '@google\u002Fgenai';\nimport fs from 'fs';\n\nconst ai = new GoogleGenerativeAI(process.env.GOOGLE_AI_API_KEY);\n\n\u002F\u002F Create store\nconst fileStore = await ai.fileSearchStores.create({\n  config: { displayName: 'my-knowledge-base' }\n});\n\n\u002F\u002F Upload document\nconst operation = await ai.fileSearchStores.uploadToFileSearchStore({\n  name: fileStore.name,\n  file: fs.createReadStream('.\u002Fmanual.pdf'),\n  config: {\n    displayName: 'Installation Manual',\n    chunkingConfig: {\n      whiteSpaceConfig: {\n        maxTokensPerChunk: 500,\n        maxOverlapTokens: 50\n      }\n    }\n  }\n});\n\n\u002F\u002F Poll until done\nwhile (!operation.done) {\n  await new Promise(resolve => setTimeout(resolve, 1000));\n  operation = await ai.operations.get({ name: operation.name });\n}\n\n\u002F\u002F Query documents\nconst model = ai.getGenerativeModel({\n  model: 'gemini-2.5-pro',  \u002F\u002F Only 2.5 Pro\u002FFlash supported\n  tools: [{\n    fileSearchTool: {\n      fileSearchStores: [fileStore.name]\n    }\n  }]\n});\n\nconst result = await model.generateContent('How do I install the product?');\nconsole.log(result.response.text());\n\n\u002F\u002F Get citations\nconst grounding = result.response.candidates[0].groundingMetadata;\nif (grounding) {\n  console.log('Sources:', grounding.groundingChunks);\n}\n```\n\n**Load `references\u002Fsetup-guide.md` for complete walkthrough with batch uploads, error handling, and production checklist.**\n\n---\n\n## Critical Rules\n\n### Always Do\n\n1. **Use delete + re-upload** for updates (documents are immutable)\n2. **Calculate 3x storage** (embeddings + metadata = ~3x file size)\n3. **Configure chunking** (500 tokens for technical docs, 800 for prose)\n4. **Poll operations** until `done: true` (with timeout)\n5. **Use force: true** when deleting stores with documents\n6. **Use Gemini 2.5 models** only (2.5-pro or 2.5-flash)\n7. **Keep metadata under 20 fields** per document\n8. **Estimate indexing costs** ($0.15\u002F1M tokens one-time)\n\n### Never Do\n\n1. **Never try to update** documents (no PATCH API exists)\n2. **Never assume storage = file size** (it's 3x)\n3. **Never skip chunking config** (defaults may not be optimal)\n4. **Never upload without polling** (operation may still be processing)\n5. **Never delete without force** if store has documents\n6. **Never use Gemini 1.5 models** (File Search requires 2.5)\n7. **Never exceed 20 metadata fields** (hard limit)\n8. **Never upload large files without cost estimate**\n\n---\n\n## Top 3 Errors Prevented\n\n### Error 1: Document Immutability\n\n**Problem:** Trying to update existing document\n\n**Solution:** Delete + re-upload pattern\n\n```typescript\n\u002F\u002F Find and delete old version\nconst docs = await ai.fileSearchStores.documents.list({\n  parent: fileStore.name\n});\nconst oldDoc = docs.documents.find(d => d.displayName === 'manual.pdf');\nif (oldDoc) {\n  await ai.fileSearchStores.documents.delete({\n    name: oldDoc.name,\n    force: true\n  });\n}\n\n\u002F\u002F Upload new version\nawait ai.fileSearchStores.uploadToFileSearchStore({\n  name: fileStore.name,\n  file: fs.createReadStream('manual-v2.pdf'),\n  config: { displayName: 'manual.pdf' }\n});\n```\n\n### Error 2: Storage Quota Exceeded\n\n**Problem:** Storage calculation wrong (3x multiplier)\n\n**Solution:** Estimate before upload\n\n```typescript\nconst fileSize = fs.statSync('data.pdf').size;\nconst estimatedStorage = fileSize * 3;  \u002F\u002F Embeddings + metadata\n\nif (estimatedStorage > 1e9) {\n  console.warn('⚠️ May exceed free tier 1 GB limit');\n}\n```\n\n### Error 3: Model Compatibility\n\n**Problem:** Using wrong model version\n\n**Solution:** Use Gemini 2.5 only\n\n```typescript\n\u002F\u002F ✅ CORRECT\nconst model = ai.getGenerativeModel({\n  model: 'gemini-2.5-pro',  \u002F\u002F or gemini-2.5-flash\n  tools: [{ fileSearchTool: { fileSearchStores: [storeName] } }]\n});\n\n\u002F\u002F ❌ WRONG\nconst model = ai.getGenerativeModel({\n  model: 'gemini-1.5-pro',  \u002F\u002F Not supported!\n  tools: [{ fileSearchTool: { fileSearchStores: [storeName] } }]\n});\n```\n\n**Load `references\u002Ferror-catalog.md` for all 8 errors with detailed solutions including chunking, operation polling, metadata limits, and force delete requirements.**\n\n---\n\n## When to Use File Search\n\n### Use File Search When:\n\n- Want fully managed RAG (no vector DB)\n- Cost predictability matters (one-time indexing)\n- Need 100+ file format support\n- Citations are important (built-in grounding)\n- Simple deployment is priority\n- Documents are relatively static\n\n### Use Alternatives When:\n\n**Cloudflare Vectorize** - Global edge performance, custom embeddings, real-time R2 updates\n**OpenAI Files API** - Assistants API, conversational threads, very large collections (10,000+)\n\n---\n\n## Common Patterns\n\n### Pattern 1: Customer Support Knowledge Base\n\n```typescript\n\u002F\u002F Upload support docs with metadata\nawait ai.fileSearchStores.uploadToFileSearchStore({\n  name: fileStore.name,\n  file: fs.createReadStream('troubleshooting.pdf'),\n  config: {\n    displayName: 'Troubleshooting Guide',\n    customMetadata: {\n      doc_type: 'support',\n      category: 'troubleshooting',\n      language: 'en'\n    }\n  }\n});\n```\n\n### Pattern 2: Batch Document Upload\n\n```typescript\nconst files = ['doc1.pdf', 'doc2.md', 'doc3.docx'];\nconst uploadPromises = files.map(file =>\n  ai.fileSearchStores.uploadToFileSearchStore({\n    name: fileStore.name,\n    file: fs.createReadStream(file),\n    config: { displayName: file }\n  })\n);\nconst operations = await Promise.all(uploadPromises);\n\n\u002F\u002F Poll all operations\nfor (const op of operations) {\n  let operation = op;\n  while (!operation.done) {\n    await new Promise(resolve => setTimeout(resolve, 1000));\n    operation = await ai.operations.get({ name: operation.name });\n  }\n  console.log('✅', operation.response.displayName);\n}\n```\n\n### Pattern 3: Document Update Flow\n\n```typescript\n\u002F\u002F 1. List existing documents\nconst docs = await ai.fileSearchStores.documents.list({\n  parent: fileStore.name\n});\n\n\u002F\u002F 2. Delete old version\nconst oldDoc = docs.documents.find(d => d.displayName === 'manual.pdf');\nif (oldDoc) {\n  await ai.fileSearchStores.documents.delete({\n    name: oldDoc.name,\n    force: true\n  });\n}\n\n\u002F\u002F 3. Upload new version\nconst operation = await ai.fileSearchStores.uploadToFileSearchStore({\n  name: fileStore.name,\n  file: fs.createReadStream('manual-v2.pdf'),\n  config: {\n    displayName: 'manual.pdf',\n    customMetadata: {\n      version: '2.0',\n      updated_at: new Date().toISOString()\n    }\n  }\n});\n\n\u002F\u002F 4. Poll until done\nwhile (!operation.done) {\n  await new Promise(resolve => setTimeout(resolve, 1000));\n  operation = await ai.operations.get({ name: operation.name });\n}\n```\n\n**Load `references\u002Fsetup-guide.md` for additional patterns including code documentation search and internal knowledge bases.**\n\n---\n\n## When to Load References\n\n### Load `references\u002Fsetup-guide.md` when:\n- First-time File Search setup\n- Need step-by-step walkthrough with all configuration options\n- Configuring batch upload strategies\n- Production deployment checklist\n- Complete API initialization patterns\n\n### Load `references\u002Ferror-catalog.md` when:\n- Encountering any of 8 common errors\n- Need detailed error solutions with code examples\n- Prevention checklist required\n- Troubleshooting upload\u002Fquery issues\n- Understanding chunking, metadata, or cost calculation problems\n\n---\n\n## Supported File Formats\n\n**100+ formats including:**\n- **Documents**: PDF, Word (.docx), Excel (.xlsx), PowerPoint (.pptx)\n- **Text**: Markdown (.md), Plain text (.txt), JSON, CSV\n- **Code**: Python, JavaScript, TypeScript, Java, C++, Go, Rust, etc.\n\n**Not supported:** Images in PDFs (text extraction only), Audio files, Video files\n\n---\n\n## Pricing\n\n**Indexing (one-time):** $0.15 per 1M tokens\n**Storage:** Free (10 GB - 1 TB depending on tier)\n**Query embeddings:** Free (retrieved context counts as input tokens)\n\n**Example:** 1,000-page document ≈ 500k tokens → Indexing cost: $0.075 → Storage: ~1.5 GB (3x multiplier)\n\n---\n\n## Chunking Guidelines\n\n**Technical docs:** 500 tokens\u002Fchunk, 50 overlap\n**Prose:** 800 tokens\u002Fchunk, 80 overlap\n**Legal:** 300 tokens\u002Fchunk, 30 overlap\n\n```typescript\nchunkingConfig: {\n  whiteSpaceConfig: {\n    maxTokensPerChunk: 500,  \u002F\u002F Smaller = more precise\n    maxOverlapTokens: 50     \u002F\u002F 10% overlap recommended\n  }\n}\n```\n\n---\n\n## Resources\n\n**References** (`references\u002F`):\n- `setup-guide.md` - Complete setup walkthrough (authentication, store creation, file upload, batch patterns, production checklist)\n- `error-catalog.md` - All 8 documented errors with solutions (immutability, storage, chunking, metadata, costs, polling, force delete, model compatibility)\n\n**Official Documentation**:\n- **File Search Overview**: https:\u002F\u002Fai.google.dev\u002Fapi\u002Ffile-search\n- **API Reference**: https:\u002F\u002Fai.google.dev\u002Fapi\u002Ffile-search\u002Fdocuments\n- **Blog Post**: https:\u002F\u002Fblog.google\u002Ftechnology\u002Fdevelopers\u002Ffile-search-gemini-api\u002F\n\n---\n\n**Questions? Issues?**\n\n1. Check `references\u002Fsetup-guide.md` for complete setup\n2. Review `references\u002Ferror-catalog.md` for all 8 errors\n3. Verify model version (must be Gemini 2.5)\n4. Check storage calculation (3x file size)\n",{"name":5,"license":15,"metadata":24,"description":7},{"version":25,"last_verified":26,"token_savings":27,"errors_prevented":28,"production_tested":29,"templates_included":14,"references_included":30},"2.0.0","2025-11-18","~65%",8,true,2,"main","c475fc64e94e803b50d14a791dfee8853fbb2c5c","integrations\u002Fclaude-skills-secondsky\u002Fplugins\u002Fgoogle-gemini-file-search\u002Fskills\u002Fgoogle-gemini-file-search","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory","https:\u002F\u002Fgithub.com\u002FElogic360","2026-11-08 22:21:22",{"source":8,"skillName":5,"commands":38,"prompt":48,"zipAvailable":29},[39,42,45],{"tool":40,"command":41},"npx","npx skills add Elogic360\u002Fsoftware-factory --skill google-gemini-file-search",{"tool":43,"command":44},"bunx","bunx skills add Elogic360\u002Fsoftware-factory --skill google-gemini-file-search",{"tool":46,"command":47},"pnpm","pnpm dlx skills add Elogic360\u002Fsoftware-factory --skill google-gemini-file-search","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fintegrations\u002Fclaude-skills-secondsky\u002Fplugins\u002Fgoogle-gemini-file-search\u002Fskills\u002Fgoogle-gemini-file-search\u002FSKILL.md",[50,56,60,64,70,74],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"references\u002Ferror-catalog.md","REFERENCE","参考文档",4992,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fintegrations\u002Fclaude-skills-secondsky\u002Fplugins\u002Fgoogle-gemini-file-search\u002Fskills\u002Fgoogle-gemini-file-search\u002Freferences\u002Ferror-catalog.md",{"path":57,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":58,"sourceHtmlUrl":59},"references\u002FREADME.md",2012,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fintegrations\u002Fclaude-skills-secondsky\u002Fplugins\u002Fgoogle-gemini-file-search\u002Fskills\u002Fgoogle-gemini-file-search\u002Freferences\u002FREADME.md",{"path":61,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":62,"sourceHtmlUrl":63},"references\u002Fsetup-guide.md",4292,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fintegrations\u002Fclaude-skills-secondsky\u002Fplugins\u002Fgoogle-gemini-file-search\u002Fskills\u002Fgoogle-gemini-file-search\u002Freferences\u002Fsetup-guide.md",{"path":65,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":68,"sourceHtmlUrl":69},"scripts\u002Fcreate-store.ts","SCRIPT","脚本",3819,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fintegrations\u002Fclaude-skills-secondsky\u002Fplugins\u002Fgoogle-gemini-file-search\u002Fskills\u002Fgoogle-gemini-file-search\u002Fscripts\u002Fcreate-store.ts",{"path":71,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":72,"sourceHtmlUrl":73},"scripts\u002FREADME.md",2201,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fintegrations\u002Fclaude-skills-secondsky\u002Fplugins\u002Fgoogle-gemini-file-search\u002Fskills\u002Fgoogle-gemini-file-search\u002Fscripts\u002FREADME.md",{"path":75,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":76,"sourceHtmlUrl":77},"templates\u002FREADME.md",2154,"https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fintegrations\u002Fclaude-skills-secondsky\u002Fplugins\u002Fgoogle-gemini-file-search\u002Fskills\u002Fgoogle-gemini-file-search\u002Ftemplates\u002FREADME.md",[79,87,94,101,108,115],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":19},"01M79G206MT3J6S0GY3C45P2KV","vss-summarize-video","Use to summarize a recorded video via the LVS summarization microservice (HITL-gated) with a VLM fallback. Not for report generation or live RTSP captioning.","Apache-2.0","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-summarize-video\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":19},"01M79G1WN63PGRE8CKD47WEBQ5","vss-setup-video-analytics-api","Use to deploy the vss-video-analytics-api REST service standalone (config-source, data-log bind, Elasticsearch, optional Kafka). Not for full warehouse deploy.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-setup-video-analytics-api\u002FSKILL.md",[],[],{"skillNo":95,"name":96,"nameZh":6,"description":97,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":98,"categories":99,"keywords":100,"publishedAt":19},"01M79G1S9F7VY3EF6H53PRP5AN","vss-setup-behavior-analytics","Use to deploy the vss-behavior-analytics service standalone (entrypoint, config-source, optional calibration). Not for the full warehouse deploy.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-setup-behavior-analytics\u002FSKILL.md",[],[],{"skillNo":102,"name":103,"nameZh":6,"description":104,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":107,"publishedAt":19},"01M79G1N1S1Y27PFZPYMNK1MTQ","vss-search-archive","Use this skill to run top-level VSS fusion search on archived video, or to ingest video files \u002F RTSP streams for search. Do NOT use for ad-hoc visual Q&A (use vss-ask-video), live captioning (use vss-deploy-dense-captioning), or video summarization and reports (use vss-summarize-video).","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-search-archive\u002FSKILL.md",[],[],{"skillNo":109,"name":110,"nameZh":6,"description":111,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":19},"01M79G1H6WN0Q6C0VCAXCZX6B7","vss-query-analytics","Use this skill when reading video-analytics metrics, incidents, alerts, and sensor data via the VA-MCP server (port 9901). Not for live VLM or incident-range narrative reports.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-query-analytics\u002FSKILL.md",[],[],{"skillNo":116,"name":117,"nameZh":6,"description":118,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":83,"licenseFileUrl":6,"sourceHtmlUrl":119,"categories":120,"keywords":121,"publishedAt":19},"01M79G1DYPNXQS7QHZ0AAF5FX5","vss-manage-video-io-storage","Use to call the VIOS REST API (sensor list, timelines, clip extraction, snapshots, add\u002Fdelete sensors and streams). Not for VLM inference or search.","https:\u002F\u002Fgithub.com\u002FElogic360\u002Fsoftware-factory\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fvss-manage-video-io-storage\u002FSKILL.md",[],[]]