[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7FDFSAZBG1FYX60C5S67RTY":3,"skill-same:01M7FDFSAZBG1FYX60C5S67RTY":62},{"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":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":6,"ownerHtmlUrl":32,"ownerBlogUrl":6,"ownerDescription":33,"updatedAt":34,"install":35,"resources":47},"01M7FDFSAZBG1FYX60C5S67RTY","knowledge-and-ocr",null,"Operate Yuxi knowledge-base, retrieval, document parsing, OCR, graph, mindmap, and knowledge evaluation workflows safely.","lza6\u002FArmor-piercing-system-prompts","lza6","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128880206?v=4","01M7EB3K7WPX5T28EJ2NVHE6EG","Chinese-tingfeng","01M7EB3K97E1KQ16QWKJMFNGX8",11,"MIT","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fyuxi\u002Fsub-skills\u002Fknowledge-and-ocr\u002FSKILL.md",[],[],"2026-11-16 10:05:34",false,"---\nname: knowledge-and-ocr\ndescription: \"Operate Yuxi knowledge-base, retrieval, document parsing, OCR,\n  graph, mindmap, and knowledge evaluation workflows safely.\"\nmetadata:\n  disco-role: operating\ndisable-model-invocation: true\nlicense: MIT\n---\n\n# Yuxi knowledge and OCR operating guide\n\nLoad this sub-skill when the task touches Yuxi knowledge bases, RAG retrieval, document upload\u002Fparse\u002Fchunk\u002Findex flows, OCR configuration, graph\u002Fmindmap features, knowledge evaluation, KB agent tools, or `read_file` OCR fallback behavior.\n\n## Owned capabilities\n\n- Knowledge-base types: Milvus document KBs, Dify Dataset read-only retrieval connectors, and Notion Data Source read-only retrieval connectors.\n- Document lifecycle: upload\u002Fimport\u002Ffetch, file records, parsing to Markdown, chunking\u002Findexing, re-indexing, file search\u002Fopen\u002Ffind\u002Fdownload, previews, and image proxy handling.\n- Retrieval: vector, keyword\u002FBM25, hybrid, reranker, file-name scoped search, and optional graph-enhanced retrieval for Milvus KBs.\n- OCR: default engine resolution, per-file engine selection, config\u002Fhealth endpoints, local\u002Fservice\u002Fcloud engines, and agent `ocr_parse_file` fallback.\n- Knowledge structure: mindmap generation\u002Fdiff\u002Fincremental update, generated sample questions, Milvus knowledge graph build\u002Fquery\u002Freset\u002Freconcile.\n- Evaluation: JSONL benchmark upload, LLM-generated benchmarks, RAG evaluation runs, result inspection, and metrics interpretation.\n\nUse sibling sub-skills instead for deployment startup details, CLI command syntax, agent runtime internals unrelated to KB data, or general repo development workflow.\n\n## Safety gates before acting\n\n1. **Classify the runtime need.** Parser facade and metadata checks are CPU-safe; real KB ingestion\u002Fretrieval requires running services such as API, Postgres, Redis\u002Ftask workers, MinIO, and Milvus. Graph build also needs Neo4j and a configured LLM provider. Some OCR engines require GPU services or cloud credentials.\n2. **Do not assume credentials.** Dify, Notion, MinerU Official, DeepSeek OCR, PaddleOCR API, rerankers, embeddings, and LLM-based graph\u002Fevaluation features may call external services. Confirm credentials and side effects before running them.\n3. **Respect file boundaries.** Knowledge files should enter via Yuxi upload\u002Fimport\u002Ffetch APIs and be stored in MinIO before document record creation. Agent OCR fallback only accepts sandbox virtual paths under user-data `workspace`, `uploads`, or `outputs` and writes Markdown under `outputs\u002Focr`.\n4. **Do not treat read-only connectors as Milvus.** Dify and Notion are normal KB types for retrieval, but public document upload\u002Fopen\u002Ffind\u002Fdownload workflows are gated to document-supporting KBs; expect read-only connectors to reject file operations.\n5. **Keep secrets out of logs and artifacts.** Config APIs intentionally redact sensitive values; never print tokens or API keys while debugging.\n\n## First routing question\n\nChoose the narrow workflow before touching code or services:\n\n- **Create\u002Fquery\u002Fmanage a KB:** use [references\u002Fknowledge-workflows.md](references\u002Fknowledge-workflows.md#knowledge-base-lifecycle-and-entry-points).\n- **OCR or parser issue:** use [references\u002Focr-engine-matrix.md](references\u002Focr-engine-matrix.md) first, then [references\u002Ftroubleshooting.md](references\u002Ftroubleshooting.md#ocr-and-parser-failures).\n- **Agent KB tools:** use [references\u002Fknowledge-workflows.md](references\u002Fknowledge-workflows.md#agent-tool-entry-points).\n- **Graph or mindmap:** use [references\u002Fknowledge-workflows.md](references\u002Fknowledge-workflows.md#mindmap-sample-questions-and-graph-workflows).\n- **Knowledge evaluation:** use [references\u002Fknowledge-workflows.md](references\u002Fknowledge-workflows.md#knowledge-evaluation-workflow).\n- **Unknown failure:** use [references\u002Ftroubleshooting.md](references\u002Ftroubleshooting.md) and classify by KB type, file status, OCR engine, service dependency, and credential requirement.\n\n## Safe entry points\n\nPrefer these entry points because they preserve Yuxi's permission, config, storage, and task semantics:\n\n- Web\u002FAPI routes under `\u002Fapi\u002Fknowledge\u002F*`, `\u002Fapi\u002Fsystem\u002Focr\u002F*`, `\u002Fapi\u002Fsystem\u002Fconfig*`, `\u002Fapi\u002Fsystem\u002Fconfig\u002Foptions*`, `\u002Fapi\u002Fevaluation\u002F*`, and `\u002Fapi\u002Fgraph\u002F*`.\n- Python service\u002Ffacade calls: `yuxi.knowledge.runtime.knowledge_base`, `yuxi.services.ocr_service.parse_document`, `yuxi.services.ocr_service.get_ocr_options`, `yuxi.services.ocr_service.check_all_ocr_health`, and `yuxi.knowledge.graphs.milvus_graph_service.MilvusGraphService`.\n- Agent tools made available by the built-in `knowledge-base` skill dependency: `list_kbs`, `get_mindmap`, `query_kb`, `search_file`, `open_kb_document`, `find_kb_document`, and `download_kb_file`.\n- Agent built-in OCR helper: `ocr_parse_file` for sandbox PDF\u002FOffice\u002Fimage to Markdown conversion.\n\nAvoid calling low-level parser functions directly unless the task is specifically parser internals. The business entry point is `parse_document`; low-level PDF\u002Fimage parsers expect an already resolved `ocr_engine` and can bypass DB\u002Fenvironment fallback if called incorrectly.\n\n## Verification candidates\n\n- CPU-safe native candidate: `pytest backend\u002Ftest\u002Funit\u002Fknowledge\u002Ftest_parser_facade.py` — proves parser registry metadata, engine selection, parser facade, credential-safe cache keys, and common document conversion behavior.\n- Service-required candidate: `pytest backend\u002Ftest\u002Fe2e\u002Ftest_ocr_config_center_e2e.py` — proves configured default OCR drives real temporary attachment parsing when API services and MinIO are available.\n- Service\u002Fmodel-required candidate: `pytest backend\u002Ftest\u002Fe2e\u002Ftest_read_file_multimodal_e2e.py` — proves image `read_file` and non-vision OCR fallback only when model\u002FOCR prerequisites are configured.\n\nRun service-required candidates only after the deployment sub-skill's runtime checks pass and the needed credentials\u002Fproviders are approved.\n","# Yuxi knowledge and OCR operating guide\n\nLoad this sub-skill when the task touches Yuxi knowledge bases, RAG retrieval, document upload\u002Fparse\u002Fchunk\u002Findex flows, OCR configuration, graph\u002Fmindmap features, knowledge evaluation, KB agent tools, or `read_file` OCR fallback behavior.\n\n## Owned capabilities\n\n- Knowledge-base types: Milvus document KBs, Dify Dataset read-only retrieval connectors, and Notion Data Source read-only retrieval connectors.\n- Document lifecycle: upload\u002Fimport\u002Ffetch, file records, parsing to Markdown, chunking\u002Findexing, re-indexing, file search\u002Fopen\u002Ffind\u002Fdownload, previews, and image proxy handling.\n- Retrieval: vector, keyword\u002FBM25, hybrid, reranker, file-name scoped search, and optional graph-enhanced retrieval for Milvus KBs.\n- OCR: default engine resolution, per-file engine selection, config\u002Fhealth endpoints, local\u002Fservice\u002Fcloud engines, and agent `ocr_parse_file` fallback.\n- Knowledge structure: mindmap generation\u002Fdiff\u002Fincremental update, generated sample questions, Milvus knowledge graph build\u002Fquery\u002Freset\u002Freconcile.\n- Evaluation: JSONL benchmark upload, LLM-generated benchmarks, RAG evaluation runs, result inspection, and metrics interpretation.\n\nUse sibling sub-skills instead for deployment startup details, CLI command syntax, agent runtime internals unrelated to KB data, or general repo development workflow.\n\n## Safety gates before acting\n\n1. **Classify the runtime need.** Parser facade and metadata checks are CPU-safe; real KB ingestion\u002Fretrieval requires running services such as API, Postgres, Redis\u002Ftask workers, MinIO, and Milvus. Graph build also needs Neo4j and a configured LLM provider. Some OCR engines require GPU services or cloud credentials.\n2. **Do not assume credentials.** Dify, Notion, MinerU Official, DeepSeek OCR, PaddleOCR API, rerankers, embeddings, and LLM-based graph\u002Fevaluation features may call external services. Confirm credentials and side effects before running them.\n3. **Respect file boundaries.** Knowledge files should enter via Yuxi upload\u002Fimport\u002Ffetch APIs and be stored in MinIO before document record creation. Agent OCR fallback only accepts sandbox virtual paths under user-data `workspace`, `uploads`, or `outputs` and writes Markdown under `outputs\u002Focr`.\n4. **Do not treat read-only connectors as Milvus.** Dify and Notion are normal KB types for retrieval, but public document upload\u002Fopen\u002Ffind\u002Fdownload workflows are gated to document-supporting KBs; expect read-only connectors to reject file operations.\n5. **Keep secrets out of logs and artifacts.** Config APIs intentionally redact sensitive values; never print tokens or API keys while debugging.\n\n## First routing question\n\nChoose the narrow workflow before touching code or services:\n\n- **Create\u002Fquery\u002Fmanage a KB:** use [references\u002Fknowledge-workflows.md](references\u002Fknowledge-workflows.md#knowledge-base-lifecycle-and-entry-points).\n- **OCR or parser issue:** use [references\u002Focr-engine-matrix.md](references\u002Focr-engine-matrix.md) first, then [references\u002Ftroubleshooting.md](references\u002Ftroubleshooting.md#ocr-and-parser-failures).\n- **Agent KB tools:** use [references\u002Fknowledge-workflows.md](references\u002Fknowledge-workflows.md#agent-tool-entry-points).\n- **Graph or mindmap:** use [references\u002Fknowledge-workflows.md](references\u002Fknowledge-workflows.md#mindmap-sample-questions-and-graph-workflows).\n- **Knowledge evaluation:** use [references\u002Fknowledge-workflows.md](references\u002Fknowledge-workflows.md#knowledge-evaluation-workflow).\n- **Unknown failure:** use [references\u002Ftroubleshooting.md](references\u002Ftroubleshooting.md) and classify by KB type, file status, OCR engine, service dependency, and credential requirement.\n\n## Safe entry points\n\nPrefer these entry points because they preserve Yuxi's permission, config, storage, and task semantics:\n\n- Web\u002FAPI routes under `\u002Fapi\u002Fknowledge\u002F*`, `\u002Fapi\u002Fsystem\u002Focr\u002F*`, `\u002Fapi\u002Fsystem\u002Fconfig*`, `\u002Fapi\u002Fsystem\u002Fconfig\u002Foptions*`, `\u002Fapi\u002Fevaluation\u002F*`, and `\u002Fapi\u002Fgraph\u002F*`.\n- Python service\u002Ffacade calls: `yuxi.knowledge.runtime.knowledge_base`, `yuxi.services.ocr_service.parse_document`, `yuxi.services.ocr_service.get_ocr_options`, `yuxi.services.ocr_service.check_all_ocr_health`, and `yuxi.knowledge.graphs.milvus_graph_service.MilvusGraphService`.\n- Agent tools made available by the built-in `knowledge-base` skill dependency: `list_kbs`, `get_mindmap`, `query_kb`, `search_file`, `open_kb_document`, `find_kb_document`, and `download_kb_file`.\n- Agent built-in OCR helper: `ocr_parse_file` for sandbox PDF\u002FOffice\u002Fimage to Markdown conversion.\n\nAvoid calling low-level parser functions directly unless the task is specifically parser internals. The business entry point is `parse_document`; low-level PDF\u002Fimage parsers expect an already resolved `ocr_engine` and can bypass DB\u002Fenvironment fallback if called incorrectly.\n\n## Verification candidates\n\n- CPU-safe native candidate: `pytest backend\u002Ftest\u002Funit\u002Fknowledge\u002Ftest_parser_facade.py` — proves parser registry metadata, engine selection, parser facade, credential-safe cache keys, and common document conversion behavior.\n- Service-required candidate: `pytest backend\u002Ftest\u002Fe2e\u002Ftest_ocr_config_center_e2e.py` — proves configured default OCR drives real temporary attachment parsing when API services and MinIO are available.\n- Service\u002Fmodel-required candidate: `pytest backend\u002Ftest\u002Fe2e\u002Ftest_read_file_multimodal_e2e.py` — proves image `read_file` and non-vision OCR fallback only when model\u002FOCR prerequisites are configured.\n\nRun service-required candidates only after the deployment sub-skill's runtime checks pass and the needed credentials\u002Fproviders are approved.\n",{"name":5,"license":15,"metadata":24,"description":7,"disable-model-invocation":26},{"disco-role":25},"operating",true,"main","d69096e239f2a89aabc2cd078c47664a9d3b1e11","06-前端与设计\u002Fyuxi\u002Fsub-skills\u002Fknowledge-and-ocr",3,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts","https:\u002F\u002Fgithub.com\u002Flza6","个人vx：Tf00798\r\n个人✈：@TFGY999","2026-10-08 14:29:22",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":26},[37,40,43],{"tool":38,"command":39},"npx","npx skills add lza6\u002FArmor-piercing-system-prompts --skill knowledge-and-ocr",{"tool":41,"command":42},"bunx","bunx skills add lza6\u002FArmor-piercing-system-prompts --skill knowledge-and-ocr",{"tool":44,"command":45},"pnpm","pnpm dlx skills add lza6\u002FArmor-piercing-system-prompts --skill knowledge-and-ocr","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fyuxi\u002Fsub-skills\u002Fknowledge-and-ocr\u002FSKILL.md",[48,54,58],{"path":49,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":52,"sourceHtmlUrl":53},"references\u002Fknowledge-workflows.md","REFERENCE","参考文档",16030,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fyuxi\u002Fsub-skills\u002Fknowledge-and-ocr\u002Freferences\u002Fknowledge-workflows.md",{"path":55,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":56,"sourceHtmlUrl":57},"references\u002Focr-engine-matrix.md",9907,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fyuxi\u002Fsub-skills\u002Fknowledge-and-ocr\u002Freferences\u002Focr-engine-matrix.md",{"path":59,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":60,"sourceHtmlUrl":61},"references\u002Ftroubleshooting.md",12559,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fyuxi\u002Fsub-skills\u002Fknowledge-and-ocr\u002Freferences\u002Ftroubleshooting.md",[63,70,75,82,89,96],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":19},"01M7FFG485KY83TD6HSC4YPCMC","obsidian-clipper-template-creator","Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator\u002FSKILL.md",[],[],{"skillNo":71,"name":65,"nameZh":6,"description":66,"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":19},"01M7FFG0T6AZ0STEJPSB3CX84K","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator-antigravity-awesome-skills-main\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":19},"01M7FFFY0QEM14SKWPDYGMXM2P","python-api","Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP\u002FmAP semantics, and source-style import diagnostics.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Fpython-api\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":19},"01M7FFFVZQZ2GJ0C0TGP7JHD20","file-evaluation","Folder-based PASCAL VOC AP\u002FmAP evaluation from ground-truth and detection text files.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Ffile-evaluation\u002FSKILL.md",[],[],{"skillNo":90,"name":91,"nameZh":6,"description":92,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":93,"categories":94,"keywords":95,"publishedAt":19},"01M7FFFSFNYSE77SJ6PE1CK2DA","object-detection-metrics","Guides legacy Object-Detection-Metrics PASCAL VOC AP\u002FmAP evaluation from detection text folders and direct Python API classes.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002FSKILL.md",[],[],{"skillNo":97,"name":98,"nameZh":6,"description":99,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":100,"categories":101,"keywords":102,"publishedAt":19},"01M7FFFP0GCK1RVXBCPQ8APBMT","nx-workspace-patterns","Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fnx-workspace-patterns\u002FSKILL.md",[],[]]