[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6HXTFDMFBCSPB16HWYANGPJ":3,"skill-same:01M6HXTFDMFBCSPB16HWYANGPJ":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":29,"linkOnly":30,"markdownContent":31,"markdownBody":32,"frontMatter":33,"refName":34,"commitSha":35,"skillDirectory":36,"forksCount":37,"repositoryHtmlUrl":38,"repositoryHomepageUrl":6,"ownerHtmlUrl":39,"ownerBlogUrl":40,"ownerDescription":6,"updatedAt":41,"install":42,"resources":55},"01M6HXTFDMFBCSPB16HWYANGPJ","prioritize-reference-files-6e4111",null,"Always read and use provided reference files for data before attempting external searches or fabricating information","HKUDS\u002FAgentPilot","HKUDS","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F118165258?v=4","01M6HW8VRNZP64G0P4MVJW2KKQ","✨Data Intelligence Lab@HKU✨","01M6HW8VWZVHBRHYH3T330PFWV",8029,"MIT","https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fopenspace\u002Flegacy\u002Fv1\u002Fgdpval_bench\u002Fskills\u002Fprioritize-reference-files-6e4111\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28],"prioritize","reference","files","prioritize reference files 6e4111","prioritize reference files","data","read","use","information","for","2026-11-10 10:12:30",false,"---\nname: prioritize-reference-files-6e4111\ndescription: Always read and use provided reference files for data before attempting external searches or fabricating information\n---\n\n# Prioritize Reference Files\n\nThis skill ensures agents correctly prioritize provided reference files over external data sources, preventing the use of fabricated or incorrect information when structured data is already available.\n\n## Core Principle\n\n**When reference files are provided in task context, you MUST read and prioritize them for data before attempting web searches or generating synthetic data.**\n\n## Step-by-Step Instructions\n\n### Step 1: Scan Task Context for Files\nBefore taking any action, identify all files provided in the task context:\n- Look for file attachments, uploads, or references\n- Check for common data formats: `.xlsx`, `.csv`, `.json`, `.pdf`, `.docx`, `.txt`\n- Note the file names and their apparent purpose\n\n### Step 2: Read Reference Files First\nRead all relevant reference files **before** any web searches:\n```\n# Example: Read provided Excel file\nfile_content = read_file(file_path=\"Massabama active listings.xlsx\", filetype=\"xlsx\")\n\n# Example: Read provided CSV\nfile_content = read_file(file_path=\"data.csv\", filetype=\"csv\")\n\n# Example: Read provided PDF\nfile_content = read_file(file_path=\"report.pdf\", filetype=\"pdf\")\n```\n\n**Important: Handling DOCX Read Failures**\n\nIf `read_file(filetype='docx')` fails with an error (common issue), use this fallback method:\n\n```bash\n# Method 1: Unzip and parse XML directly (docx is a zip archive)\nmkdir -p temp_docx && cd temp_docx\nunzip -o ..\u002Fdocument.docx\n# Extract text from word\u002Fdocument.xml\ncat word\u002Fdocument.xml | grep -oP '(?\u003C=>)[^\u003C>]+' > extracted_text.txt\n```\n\n```python\n# Method 2: Use python-docx via shell\nrun_shell(command=\"python -c \\\"from docx import Document; doc = Document('document.docx'); print('\\n'.join([p.text for p in doc.paragraphs]))\\\"\")\n```\n\n```python\n# Method 3: Use shell_agent for robust extraction\nextraction_result = shell_agent(task=\"Extract all text content from document.docx using any reliable method (unzip+XML, python-docx, or pandoc)\")\n```\n\nAfter extracting via fallback:\n1. Verify the extracted content is complete and readable\n2. Proceed to use this extracted data as your reference file content\n3. Continue with Step 3 (Extract and Validate Data) using the fallback-extracted content\n\n### Step 3: Extract and Validate Data\n- Extract structured data from the reference files\n- Verify the data is complete and relevant to your task\n- Map the file data to your required output format\n\n### Step 4: Use Reference Data as Primary Source\n- Populate outputs using data from reference files\n- Cite the reference files when using their data\n- Only mark data as \"to be verified\" if reference files are incomplete\n\n### Step 5: Search Web Only When Necessary\nWeb searches should only occur when:\n- Reference files do not contain the needed information\n- You need to verify or supplement reference file data\n- The task explicitly requires current\u002Fexternal information\n\n```python\n# Decision flow example\nif has_reference_file(\"listings.xlsx\"):\n    data = extract_from_excel(\"listings.xlsx\")\n    # Use this data for output\nelse:\n    # Only then search web\n    data = search_web(\"property listings\")\n```\n\n## Anti-Patterns to Avoid\n\n❌ **DO NOT** ignore provided files and search the web first\n❌ **DO NOT** fabricate data when reference files exist\n❌ **DO NOT** assume reference files are irrelevant without reading them\n❌ **DO NOT** create placeholder\u002Fsynthetic data before checking available files\n\n## Checklist\n\nBefore completing any data-dependent task:\n\n- [ ] I have identified all files provided in the task context\n- [ ] I have read all relevant reference files\n- [ ] I have extracted available data from reference files\n- [ ] I am using reference file data as my primary source\n- [ ] Web searches are only for gaps not covered by reference files\n- [ ] I have cited reference files in my output where applicable\n\n## Example Application\n\n**Task**: Create property listings PDF from provided data\n\n**Correct Approach**:\n1. Find `Massabama active listings.xlsx` in task context\n2. Read the Excel file to extract property addresses, prices, details\n3. Use this real data to populate the PDF\n4. Only search web if verifying map locations or supplementing missing fields\n\n**Incorrect Approach**:\n1. Ignore the Excel file\n2. Search web for \"Massabama properties\"\n3. Fabricate placeholder property data\n4. Create PDF with synthetic\u002Ffabricated information\n","# Prioritize Reference Files\n\nThis skill ensures agents correctly prioritize provided reference files over external data sources, preventing the use of fabricated or incorrect information when structured data is already available.\n\n## Core Principle\n\n**When reference files are provided in task context, you MUST read and prioritize them for data before attempting web searches or generating synthetic data.**\n\n## Step-by-Step Instructions\n\n### Step 1: Scan Task Context for Files\nBefore taking any action, identify all files provided in the task context:\n- Look for file attachments, uploads, or references\n- Check for common data formats: `.xlsx`, `.csv`, `.json`, `.pdf`, `.docx`, `.txt`\n- Note the file names and their apparent purpose\n\n### Step 2: Read Reference Files First\nRead all relevant reference files **before** any web searches:\n```\n# Example: Read provided Excel file\nfile_content = read_file(file_path=\"Massabama active listings.xlsx\", filetype=\"xlsx\")\n\n# Example: Read provided CSV\nfile_content = read_file(file_path=\"data.csv\", filetype=\"csv\")\n\n# Example: Read provided PDF\nfile_content = read_file(file_path=\"report.pdf\", filetype=\"pdf\")\n```\n\n**Important: Handling DOCX Read Failures**\n\nIf `read_file(filetype='docx')` fails with an error (common issue), use this fallback method:\n\n```bash\n# Method 1: Unzip and parse XML directly (docx is a zip archive)\nmkdir -p temp_docx && cd temp_docx\nunzip -o ..\u002Fdocument.docx\n# Extract text from word\u002Fdocument.xml\ncat word\u002Fdocument.xml | grep -oP '(?\u003C=>)[^\u003C>]+' > extracted_text.txt\n```\n\n```python\n# Method 2: Use python-docx via shell\nrun_shell(command=\"python -c \\\"from docx import Document; doc = Document('document.docx'); print('\\n'.join([p.text for p in doc.paragraphs]))\\\"\")\n```\n\n```python\n# Method 3: Use shell_agent for robust extraction\nextraction_result = shell_agent(task=\"Extract all text content from document.docx using any reliable method (unzip+XML, python-docx, or pandoc)\")\n```\n\nAfter extracting via fallback:\n1. Verify the extracted content is complete and readable\n2. Proceed to use this extracted data as your reference file content\n3. Continue with Step 3 (Extract and Validate Data) using the fallback-extracted content\n\n### Step 3: Extract and Validate Data\n- Extract structured data from the reference files\n- Verify the data is complete and relevant to your task\n- Map the file data to your required output format\n\n### Step 4: Use Reference Data as Primary Source\n- Populate outputs using data from reference files\n- Cite the reference files when using their data\n- Only mark data as \"to be verified\" if reference files are incomplete\n\n### Step 5: Search Web Only When Necessary\nWeb searches should only occur when:\n- Reference files do not contain the needed information\n- You need to verify or supplement reference file data\n- The task explicitly requires current\u002Fexternal information\n\n```python\n# Decision flow example\nif has_reference_file(\"listings.xlsx\"):\n    data = extract_from_excel(\"listings.xlsx\")\n    # Use this data for output\nelse:\n    # Only then search web\n    data = search_web(\"property listings\")\n```\n\n## Anti-Patterns to Avoid\n\n❌ **DO NOT** ignore provided files and search the web first\n❌ **DO NOT** fabricate data when reference files exist\n❌ **DO NOT** assume reference files are irrelevant without reading them\n❌ **DO NOT** create placeholder\u002Fsynthetic data before checking available files\n\n## Checklist\n\nBefore completing any data-dependent task:\n\n- [ ] I have identified all files provided in the task context\n- [ ] I have read all relevant reference files\n- [ ] I have extracted available data from reference files\n- [ ] I am using reference file data as my primary source\n- [ ] Web searches are only for gaps not covered by reference files\n- [ ] I have cited reference files in my output where applicable\n\n## Example Application\n\n**Task**: Create property listings PDF from provided data\n\n**Correct Approach**:\n1. Find `Massabama active listings.xlsx` in task context\n2. Read the Excel file to extract property addresses, prices, details\n3. Use this real data to populate the PDF\n4. Only search web if verifying map locations or supplementing missing fields\n\n**Incorrect Approach**:\n1. Ignore the Excel file\n2. Search web for \"Massabama properties\"\n3. Fabricate placeholder property data\n4. Create PDF with synthetic\u002Ffabricated information\n",{"name":5,"description":7},"main","6b84b0689ca0d9e7b1e4061188ff6b60758e0f8a","openspace\u002Flegacy\u002Fv1\u002Fgdpval_bench\u002Fskills\u002Fprioritize-reference-files-6e4111",944,"https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot","https:\u002F\u002Fgithub.com\u002FHKUDS","https:\u002F\u002Fsites.google.com\u002Fview\u002Fchaoh","2026-11-12 22:26:39",{"source":8,"skillName":5,"commands":43,"prompt":53,"zipAvailable":54},[44,47,50],{"tool":45,"command":46},"npx","npx skills add HKUDS\u002FAgentPilot --skill prioritize-reference-files-6e4111",{"tool":48,"command":49},"bunx","bunx skills add HKUDS\u002FAgentPilot --skill prioritize-reference-files-6e4111",{"tool":51,"command":52},"pnpm","pnpm dlx skills add HKUDS\u002FAgentPilot --skill prioritize-reference-files-6e4111","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fopenspace\u002Flegacy\u002Fv1\u002Fgdpval_bench\u002Fskills\u002Fprioritize-reference-files-6e4111\u002FSKILL.md",true,[56],{"path":57,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":60,"sourceHtmlUrl":61},".skill_id","OTHER","其他",47,"https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fopenspace\u002Flegacy\u002Fv1\u002Fgdpval_bench\u002Fskills\u002Fprioritize-reference-files-6e4111\u002F.skill_id",[63,82,101,126,149,171],{"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":29},"01M6HYB297NDABRERJRVG0K8R0","agentpilot","Set up AgentPilot and use its single entry to prepare missing inputs and run configured RSI experiments. Use when the user requests AgentPilot installation or work with AgentPilot.","https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fplugins\u002Fagentpilot\u002Fskills\u002Fagentpilot\u002FSKILL.md",[],[65,70,71,72,73,74,75,76,26,77,78,79,80,81],"prepare","requests","with","missing","installation","rsi","up","user","work","SET","run","experiments",{"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":29},"01M6HYAYWXTTE1J7DJM0MFNQ4H","remember","Review auto-memory entries and propose promotions to OPENSPACE.md, OPENSPACE.local.md, or retained auto-memory topics; detect duplicates, stale facts, conflicts, and ambiguous entries before making changes.","https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fopenspace\u002Fopenspace\u002Fskills\u002Fremember\u002FSKILL.md",[],[84,89,90,91,92,93,94,95,96,97,98,99,100],"memory","review","auto","openspace","changes","facts","detect","entries","local","conflicts","stale","duplicates",{"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":15,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":112,"publishedAt":29},"01M6HYAVMPM4SBDV60R2B5HPV8","skill-discovery","Search reusable OpenSpace skills locally, and browse cloud packages step by step when a cloud skill may help. Reusing proven skills saves tokens, improves reliability, and extends your capabilities beyond built-in tools.","https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fopenspace\u002Fopenspace\u002Fhost_skills\u002Fskill-discovery\u002FSKILL.md",[107],{"categoryNo":108,"categoryName":109,"taxonomyType":110,"taxonomyTypeDesc":111},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[113,114,115,92,116,117,118,119,120,121,122,123,124,125],"skill discovery","discovery","reusable","step by step","tokens","packages","help","cloud","reliability","capabilities","browse","search","Tools",{"skillNo":127,"name":128,"nameZh":6,"description":129,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":130,"categories":131,"keywords":133,"publishedAt":29},"01M6HYAR811B3RB0M50RBEDGN6","delegate-task","Delegate tasks to OpenSpace — a full-stack autonomous worker for coding, DevOps, web research, and desktop automation, backed by an extensive MCP tool and skill library. Skills auto-improve through use, reducing token consumption over time. A cloud community lets agents share and collectively evolve reusable skills.","https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fopenspace\u002Fopenspace\u002Fhost_skills\u002Fdelegate-task\u002FSKILL.md",[132],{"categoryNo":108,"categoryName":109,"taxonomyType":110,"taxonomyTypeDesc":111},[134,135,136,92,137,138,139,140,141,115,91,26,142,143,144,28,145,146,147,148],"delegate","task","delegate task","time","improve","desktop automation","worker","community","autonomous","tool","stack","share","tasks","automation","MCP",{"skillNo":150,"name":151,"nameZh":6,"description":152,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":153,"categories":154,"keywords":155,"publishedAt":29},"01M6HYAM9FSGV6E52AYENDQXYH","write-special-chars-ts-check","Workaround for write_file failures caused by special characters (apostrophes, backticks, template literals) using a Python heredoc script, plus a reliable TypeScript type-checking step that filters output to only the newly-written files to avoid being misled by pre-existing errors.","https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fopenspace\u002Flegacy\u002Fv1\u002Fshowcase\u002Fskills\u002Fwrite-special-chars-ts-check\u002FSKILL.md",[],[156,157,158,159,21,160,161,162,163,164,165,166,167,168,28,169,170],"write special chars ts check","write","check","type","filters","only","template","file","python","typescript","output","Existing","type-checking","errors","script",{"skillNo":172,"name":173,"nameZh":6,"description":174,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":175,"categories":176,"keywords":177,"publishedAt":29},"01M6HYAH05HGV7SVWSVR8PJRQD","write-special-chars-readback","Workaround for write_file failures caused by special characters (apostrophes, backticks, template literals) using a Python heredoc script, with mandatory post-write verification via read_file and cleanup of \u002Ftmp helper scripts.","https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fopenspace\u002Flegacy\u002Fv1\u002Fshowcase\u002Fskills\u002Fwrite-special-chars-readback\u002FSKILL.md",[],[178,179,157,72,28,170,180,181,182,183,25,162,163,184,185,164],"write special chars readback","readback","helper","post","via","scripts","verification","cleanup"]