[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6HY2WCG8TG2N2GQS91QTG7R":3,"skill-same:01M6HY2WCG8TG2N2GQS91QTG7R":72},{"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":23,"publishedAt":39,"linkOnly":40,"markdownContent":41,"markdownBody":42,"frontMatter":43,"refName":44,"commitSha":45,"skillDirectory":46,"forksCount":47,"repositoryHtmlUrl":48,"repositoryHomepageUrl":6,"ownerHtmlUrl":49,"ownerBlogUrl":50,"ownerDescription":6,"updatedAt":51,"install":52,"resources":65},"01M6HY2WCG8TG2N2GQS91QTG7R","context-anchored-fallback-report",null,"Generate documents with write_file when retrieval tools fail, with explicit guardrails against task context drift","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\u002Fwrite-file-fallback-report-enhanced-4bc5c2\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000034","行政与办公人员","PROFESSION","适用职业",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38],"report","context","fallback","context anchored fallback report","documents","generate","file","Guardrails","Tools","write","task","drift","retrieval","with","context drift","2026-11-10 10:12:30",false,"---\nname: context-anchored-fallback-report\ndescription: Generate documents with write_file when retrieval tools fail, with explicit guardrails against task context drift\n---\n\n# Context-Anchored Fallback Report Generation\n\n## Critical Warning: Context Drift Prevention\n\n**DO NOT ABANDON THE ORIGINAL TASK** when tools fail. The most common failure mode is switching to an unrelated task after encountering tool errors.\n\n```\nCONTEXT_DRIFT_WARNING:\n  IF tool_failures occur (read_webpage, search_web, execute_code_sandbox)\n  THEN you MUST stay on the ORIGINAL task\n  AND you MUST NOT switch to unrelated tasks\n  AND you MUST generate output for the original goal using write_file\n```\n\n## When to Use This Skill\n\nActivate this workflow when **multiple primary data source tools fail** during document\u002Freport generation:\n\n### Trigger Decision Tree\n\n```\nSTART: Attempting to generate document\u002Freport\n  │\n  ├─→ Tool #1 fails (read_webpage\u002Fsearch_web\u002Fread_file\u002Fexecute_code_sandbox)?\n  │   │\n  │   ├─→ YES → Count = 1, Continue attempting\n  │   │\n  │   └─→ NO → Proceed normally with retrieved data\n  │\n  ├─→ Tool #2 fails (same or different tool)?\n  │   │\n  │   ├─→ YES → Count = 2, TRIGGER FALLBACK NOW\n  │   │         │\n  │   │         ├─→ LOCK to original task goal (write it down)\n  │   │         ├─→ Stop attempting failing tools\n  │   │         └─→ Generate document with write_file + embedded knowledge\n  │   │\n  │   └─→ NO → Continue with 1-2 more attempts max\n  │\n  └─→ 2+ consecutive failures without document output?\n      │\n      └─→ CRITICAL ERROR: You are drifting from task. STOP and apply this skill.\n```\n\n### Tool Failure Indicators\n\n```python\nFALLBACK_TRIGGERS = {\n    \"read_file\": \"returns binary\u002Fimage data instead of text, or file not found\",\n    \"search_web\": \"returns unknown error, empty results, or rate limited\",\n    \"read_webpage\": \"returns unknown error, connection timeout, or blocked access\",\n    \"execute_code_sandbox\": \"fails with unknown error or timeout repeatedly\",\n    \"any_retrieval_tool\": \"2+ consecutive failures on data retrieval attempts\"\n}\n```\n\n**Activation threshold**: 2+ failures from the above triggers = IMMEDIATE fallback activation.\n\n## Step-by-Step Instructions\n\n### Step 0: Context Anchor (BEFORE pivoting)\n\n**CRITICAL STEP**: Before any fallback action, write down the original task:\n\n```\nCONTEXT_ANCHOR = \"[Original task goal from user request]\"\n```\n\nExample:\n```\nCONTEXT_ANCHOR = \"Create PACT Act veterans benefits document explaining claim process\"\n```\n\n**Rule**: Every subsequent action must reference and serve this CONTEXT_ANCHOR. Never deviate.\n\n### Step 1: Detect and Declare Failure Pattern\n\nWhen 2+ tool failures occur:\n\n1. **State the limitation clearly** (once, briefly):\n   ```\n   Note: External data sources (web search, webpage access) are currently unavailable.\n   This report will be generated using established domain knowledge.\n   ```\n\n2. **Declare the pivot explicitly**:\n   ```\n   Pivoting to write_file-based document generation with embedded knowledge.\n   Task remains: [CONTEXT_ANCHOR]\n   ```\n\n3. **Do NOT**:\n   - Apologize repeatedly\n   - Continue retrying failed tools beyond 2 attempts\n   - Suggest the task cannot be completed\n   - Switch to a different task\n\n### Step 2: Pre-Compute Available Knowledge\n\nBefore writing, inventory what you CAN provide:\n\n```python\n# Knowledge inventory for the task\navailable_knowledge = {\n    \"domain_frameworks\": \"General best practices, standard procedures\",\n    \"structural_templates\": \"Professional document formats for this type\",\n    \"actionable_guidance\": \"Step-by-step processes based on established patterns\",\n    \"placeholder_markers\": \"Where specific data would enhance (clearly marked)\"\n}\n```\n\n### Step 3: Generate Document with write_file\n\nCreate professionally structured content:\n\n```markdown\n# [Document Title - aligned with CONTEXT_ANCHOR]\n\n## Executive Summary\n[2-3 sentences on what this document covers, noting data source limitations if relevant]\n\n## Background & Scope\n[Context based on embedded domain knowledge]\n\n## Core Content\n[Organized sections with:\n  - Clear headers (##, ###)\n  - Bullet points and numbered lists\n  - Tables where appropriate\n  - Code blocks for technical content\n]\n\n## Data Source Notes\n> **Note**: Specific data points that would typically come from [expected sources]\n> were unavailable at generation time. Content reflects established practices\n> in this domain.\n\n## Actionable Recommendations\n1. [Concrete step 1]\n2. [Concrete step 2]\n3. [Next steps for obtaining specific data if needed]\n\n## Task Completion Status\n- **Original Goal**: [CONTEXT_ANCHOR]\n- **Completion Method**: Generated via write_file with embedded domain knowledge\n- **Data Limitations**: [Brief note on what was unavailable]\n```\n\n### Step 4: Execute and Verify\n\n```python\n# Execution pattern\nwrite_file(\n    path=\"[output_path].md\",  # or .txt, .html based on task\n    content=professionally_structured_markdown\n)\n\n# Verification\nlist_dir(path=\".\")  # Confirm file creation\n# Optional: read_file to verify content quality\n```\n\n### Step 5: Context Integrity Check\n\nBefore declaring task complete, verify:\n\n```\nCONTEXT_CHECKLIST = [\n    \"✓ Output addresses ORIGINAL task goal\",\n    \"✓ No unrelated task content included\",\n    \"✓ Document is professionally structured\",\n    \"✓ Limitations are transparent but not over-emphasized\",\n    \"✓ Actionable guidance is provided\",\n    \"✓ File was successfully created\"\n]\n```\n\nIf any check fails, revise before completing the task.\n\n## Code Example: Full Pattern\n\n```python\ndef context_anchored_fallback(original_task, failed_tools):\n    \"\"\"\n    Generate document when tools fail, maintaining task context.\n    \"\"\"\n    # Step 0: Anchor context\n    context_anchor = original_task\n    print(f\"CONTEXT ANCHOR: {context_anchor}\")\n    \n    # Step 1: Declare pivot\n    limitation_note = \"\"\"\n    Note: External data retrieval tools experienced failures.\n    This document is generated using established domain knowledge.\n    \"\"\"\n    \n    # Step 2-3: Generate structured content\n    document = f\"\"\"# {original_task} - Report\n\n## Executive Summary\n{limitation_note.strip()}\nThis report provides guidance based on established domain knowledge.\n\n## Core Guidance\n### Key Framework\n[Structured content with headers, bullets, tables]\n\n### Process Overview\n1. Step one\n2. Step two\n3. Step three\n\n## Data Source Notes\n> Specific data from [expected sources] was unavailable.\n> Recommendations reflect established best practices.\n\n## Action Items\n1. [Actionable step 1]\n2. [Actionable step 2]\n\n## Task Status\n- **Goal**: {context_anchor}\n- **Status**: Completed via fallback generation\n- **Limitations**: Data sources unavailable, content based on domain knowledge\n\"\"\"\n    \n    # Step 4: Execute\n    output_path = \"generated_report.md\"\n    write_file(path=output_path, content=document)\n    \n    # Step 5: Verify\n    files = list_dir(path=\".\")\n    assert output_path in [f['name'] for f in files]\n    \n    return f\"Task completed: {context_anchor}\"\n```\n\n## Guardrails Against Context Drift\n\n### RED FLAGS (Stop immediately if you notice these)\n\n| Red Flag | Corrective Action |\n|----------|-------------------|\n| Thinking about a different task than the original | STOP. Re-read CONTEXT_ANCHOR. |\n| More than 3 tool retry attempts | STOP. Trigger fallback immediately. |\n| Considering \"maybe this task isn't possible\" | STOP. Generate with available knowledge. |\n| Output doesn't match original task goal | STOP. Regenerate aligned with CONTEXT_ANCHOR. |\n| Spending >5 iterations on tool troubleshooting | STOP. Pivot to write_file approach. |\n\n### GREEN FLAGS (You're on track)\n\n- Every output references the original task goal\n- Tool failures trigger fallback within 2 attempts\n- Document is generated even with data limitations\n- Final output clearly serves the original user request\n- No unrelated task content appears in output\n\n## Best Practices\n\n| Do | Don't |\n|----|-------|\n| Write CONTEXT_ANCHOR before any fallback action | Start generating without anchoring to original task |\n| Trigger fallback after 2 tool failures | Retry failing tools 5+ times |\n| Generate complete document with available knowledge | Leave task incomplete due to missing data |\n| Mark unverifiable specifics clearly | Present猜测 as verified facts |\n| Verify output matches original task | Assume task is complete without checking |\n| Use professional document structure | Output unstructured text |\n\n## Common Pitfalls & Solutions\n\n| Pitfall | Solution |\n|---------|----------|\n| **Context drift after tool failures** | Write CONTEXT_ANCHOR visibly before generating; check every output against it |\n| **Over-apologizing for limitations** | State limitation once, then deliver value |\n| **Under-delivering (no output)** | A structured partial report beats no report |\n| **Misrepresenting certainty** | Use hedging: \"typically\", \"generally\", \"established practice\" |\n| **Skipping verification** | Always run list_dir to confirm file creation |\n\n## Success Criteria\n\nTask is successfully completed when ALL are true:\n\n- [ ] Document generated despite tool failures\n- [ ] Output addresses ORIGINAL task goal (CONTEXT_ANCHOR)\n- [ ] No unrelated task content in output\n- [ ] Professional structure (headers, sections, lists, tables)\n- [ ] Limitations transparently noted (not over-emphasized)\n- [ ] Actionable guidance provided\n- [ ] File successfully created and verified\n- [ ] Fallback triggered within 2-3 tool failures (not after excessive retries)\n\n## Quick Reference Card\n\n```\n┌─────────────────────────────────────────────────────────┐\n│  CONTEXT-ANCHORED FALLBACK - QUICK TRIGGER              │\n├─────────────────────────────────────────────────────────┤\n│  IF 2+ retrieval tools fail                             │\n│  THEN:                                                  │\n│    1. Write CONTEXT_ANCHOR = [original task]            │\n│    2. Stop retrying failed tools                        │\n│    3. Generate document with write_file                 │\n│    4. Use embedded domain knowledge                     │\n│    5. Verify output matches CONTEXT_ANCHOR              │\n│    6. Confirm file created with list_dir                │\n│                                                         │\n│  NEVER:                                                 │\n│    - Switch to unrelated task                           │\n│    - Retry >3 times without pivoting                    │\n│    - Abandon original task goal                         │\n└─────────────────────────────────────────────────────────┘\n```\n","# Context-Anchored Fallback Report Generation\n\n## Critical Warning: Context Drift Prevention\n\n**DO NOT ABANDON THE ORIGINAL TASK** when tools fail. The most common failure mode is switching to an unrelated task after encountering tool errors.\n\n```\nCONTEXT_DRIFT_WARNING:\n  IF tool_failures occur (read_webpage, search_web, execute_code_sandbox)\n  THEN you MUST stay on the ORIGINAL task\n  AND you MUST NOT switch to unrelated tasks\n  AND you MUST generate output for the original goal using write_file\n```\n\n## When to Use This Skill\n\nActivate this workflow when **multiple primary data source tools fail** during document\u002Freport generation:\n\n### Trigger Decision Tree\n\n```\nSTART: Attempting to generate document\u002Freport\n  │\n  ├─→ Tool #1 fails (read_webpage\u002Fsearch_web\u002Fread_file\u002Fexecute_code_sandbox)?\n  │   │\n  │   ├─→ YES → Count = 1, Continue attempting\n  │   │\n  │   └─→ NO → Proceed normally with retrieved data\n  │\n  ├─→ Tool #2 fails (same or different tool)?\n  │   │\n  │   ├─→ YES → Count = 2, TRIGGER FALLBACK NOW\n  │   │         │\n  │   │         ├─→ LOCK to original task goal (write it down)\n  │   │         ├─→ Stop attempting failing tools\n  │   │         └─→ Generate document with write_file + embedded knowledge\n  │   │\n  │   └─→ NO → Continue with 1-2 more attempts max\n  │\n  └─→ 2+ consecutive failures without document output?\n      │\n      └─→ CRITICAL ERROR: You are drifting from task. STOP and apply this skill.\n```\n\n### Tool Failure Indicators\n\n```python\nFALLBACK_TRIGGERS = {\n    \"read_file\": \"returns binary\u002Fimage data instead of text, or file not found\",\n    \"search_web\": \"returns unknown error, empty results, or rate limited\",\n    \"read_webpage\": \"returns unknown error, connection timeout, or blocked access\",\n    \"execute_code_sandbox\": \"fails with unknown error or timeout repeatedly\",\n    \"any_retrieval_tool\": \"2+ consecutive failures on data retrieval attempts\"\n}\n```\n\n**Activation threshold**: 2+ failures from the above triggers = IMMEDIATE fallback activation.\n\n## Step-by-Step Instructions\n\n### Step 0: Context Anchor (BEFORE pivoting)\n\n**CRITICAL STEP**: Before any fallback action, write down the original task:\n\n```\nCONTEXT_ANCHOR = \"[Original task goal from user request]\"\n```\n\nExample:\n```\nCONTEXT_ANCHOR = \"Create PACT Act veterans benefits document explaining claim process\"\n```\n\n**Rule**: Every subsequent action must reference and serve this CONTEXT_ANCHOR. Never deviate.\n\n### Step 1: Detect and Declare Failure Pattern\n\nWhen 2+ tool failures occur:\n\n1. **State the limitation clearly** (once, briefly):\n   ```\n   Note: External data sources (web search, webpage access) are currently unavailable.\n   This report will be generated using established domain knowledge.\n   ```\n\n2. **Declare the pivot explicitly**:\n   ```\n   Pivoting to write_file-based document generation with embedded knowledge.\n   Task remains: [CONTEXT_ANCHOR]\n   ```\n\n3. **Do NOT**:\n   - Apologize repeatedly\n   - Continue retrying failed tools beyond 2 attempts\n   - Suggest the task cannot be completed\n   - Switch to a different task\n\n### Step 2: Pre-Compute Available Knowledge\n\nBefore writing, inventory what you CAN provide:\n\n```python\n# Knowledge inventory for the task\navailable_knowledge = {\n    \"domain_frameworks\": \"General best practices, standard procedures\",\n    \"structural_templates\": \"Professional document formats for this type\",\n    \"actionable_guidance\": \"Step-by-step processes based on established patterns\",\n    \"placeholder_markers\": \"Where specific data would enhance (clearly marked)\"\n}\n```\n\n### Step 3: Generate Document with write_file\n\nCreate professionally structured content:\n\n```markdown\n# [Document Title - aligned with CONTEXT_ANCHOR]\n\n## Executive Summary\n[2-3 sentences on what this document covers, noting data source limitations if relevant]\n\n## Background & Scope\n[Context based on embedded domain knowledge]\n\n## Core Content\n[Organized sections with:\n  - Clear headers (##, ###)\n  - Bullet points and numbered lists\n  - Tables where appropriate\n  - Code blocks for technical content\n]\n\n## Data Source Notes\n> **Note**: Specific data points that would typically come from [expected sources]\n> were unavailable at generation time. Content reflects established practices\n> in this domain.\n\n## Actionable Recommendations\n1. [Concrete step 1]\n2. [Concrete step 2]\n3. [Next steps for obtaining specific data if needed]\n\n## Task Completion Status\n- **Original Goal**: [CONTEXT_ANCHOR]\n- **Completion Method**: Generated via write_file with embedded domain knowledge\n- **Data Limitations**: [Brief note on what was unavailable]\n```\n\n### Step 4: Execute and Verify\n\n```python\n# Execution pattern\nwrite_file(\n    path=\"[output_path].md\",  # or .txt, .html based on task\n    content=professionally_structured_markdown\n)\n\n# Verification\nlist_dir(path=\".\")  # Confirm file creation\n# Optional: read_file to verify content quality\n```\n\n### Step 5: Context Integrity Check\n\nBefore declaring task complete, verify:\n\n```\nCONTEXT_CHECKLIST = [\n    \"✓ Output addresses ORIGINAL task goal\",\n    \"✓ No unrelated task content included\",\n    \"✓ Document is professionally structured\",\n    \"✓ Limitations are transparent but not over-emphasized\",\n    \"✓ Actionable guidance is provided\",\n    \"✓ File was successfully created\"\n]\n```\n\nIf any check fails, revise before completing the task.\n\n## Code Example: Full Pattern\n\n```python\ndef context_anchored_fallback(original_task, failed_tools):\n    \"\"\"\n    Generate document when tools fail, maintaining task context.\n    \"\"\"\n    # Step 0: Anchor context\n    context_anchor = original_task\n    print(f\"CONTEXT ANCHOR: {context_anchor}\")\n    \n    # Step 1: Declare pivot\n    limitation_note = \"\"\"\n    Note: External data retrieval tools experienced failures.\n    This document is generated using established domain knowledge.\n    \"\"\"\n    \n    # Step 2-3: Generate structured content\n    document = f\"\"\"# {original_task} - Report\n\n## Executive Summary\n{limitation_note.strip()}\nThis report provides guidance based on established domain knowledge.\n\n## Core Guidance\n### Key Framework\n[Structured content with headers, bullets, tables]\n\n### Process Overview\n1. Step one\n2. Step two\n3. Step three\n\n## Data Source Notes\n> Specific data from [expected sources] was unavailable.\n> Recommendations reflect established best practices.\n\n## Action Items\n1. [Actionable step 1]\n2. [Actionable step 2]\n\n## Task Status\n- **Goal**: {context_anchor}\n- **Status**: Completed via fallback generation\n- **Limitations**: Data sources unavailable, content based on domain knowledge\n\"\"\"\n    \n    # Step 4: Execute\n    output_path = \"generated_report.md\"\n    write_file(path=output_path, content=document)\n    \n    # Step 5: Verify\n    files = list_dir(path=\".\")\n    assert output_path in [f['name'] for f in files]\n    \n    return f\"Task completed: {context_anchor}\"\n```\n\n## Guardrails Against Context Drift\n\n### RED FLAGS (Stop immediately if you notice these)\n\n| Red Flag | Corrective Action |\n|----------|-------------------|\n| Thinking about a different task than the original | STOP. Re-read CONTEXT_ANCHOR. |\n| More than 3 tool retry attempts | STOP. Trigger fallback immediately. |\n| Considering \"maybe this task isn't possible\" | STOP. Generate with available knowledge. |\n| Output doesn't match original task goal | STOP. Regenerate aligned with CONTEXT_ANCHOR. |\n| Spending >5 iterations on tool troubleshooting | STOP. Pivot to write_file approach. |\n\n### GREEN FLAGS (You're on track)\n\n- Every output references the original task goal\n- Tool failures trigger fallback within 2 attempts\n- Document is generated even with data limitations\n- Final output clearly serves the original user request\n- No unrelated task content appears in output\n\n## Best Practices\n\n| Do | Don't |\n|----|-------|\n| Write CONTEXT_ANCHOR before any fallback action | Start generating without anchoring to original task |\n| Trigger fallback after 2 tool failures | Retry failing tools 5+ times |\n| Generate complete document with available knowledge | Leave task incomplete due to missing data |\n| Mark unverifiable specifics clearly | Present猜测 as verified facts |\n| Verify output matches original task | Assume task is complete without checking |\n| Use professional document structure | Output unstructured text |\n\n## Common Pitfalls & Solutions\n\n| Pitfall | Solution |\n|---------|----------|\n| **Context drift after tool failures** | Write CONTEXT_ANCHOR visibly before generating; check every output against it |\n| **Over-apologizing for limitations** | State limitation once, then deliver value |\n| **Under-delivering (no output)** | A structured partial report beats no report |\n| **Misrepresenting certainty** | Use hedging: \"typically\", \"generally\", \"established practice\" |\n| **Skipping verification** | Always run list_dir to confirm file creation |\n\n## Success Criteria\n\nTask is successfully completed when ALL are true:\n\n- [ ] Document generated despite tool failures\n- [ ] Output addresses ORIGINAL task goal (CONTEXT_ANCHOR)\n- [ ] No unrelated task content in output\n- [ ] Professional structure (headers, sections, lists, tables)\n- [ ] Limitations transparently noted (not over-emphasized)\n- [ ] Actionable guidance provided\n- [ ] File successfully created and verified\n- [ ] Fallback triggered within 2-3 tool failures (not after excessive retries)\n\n## Quick Reference Card\n\n```\n┌─────────────────────────────────────────────────────────┐\n│  CONTEXT-ANCHORED FALLBACK - QUICK TRIGGER              │\n├─────────────────────────────────────────────────────────┤\n│  IF 2+ retrieval tools fail                             │\n│  THEN:                                                  │\n│    1. Write CONTEXT_ANCHOR = [original task]            │\n│    2. Stop retrying failed tools                        │\n│    3. Generate document with write_file                 │\n│    4. Use embedded domain knowledge                     │\n│    5. Verify output matches CONTEXT_ANCHOR              │\n│    6. Confirm file created with list_dir                │\n│                                                         │\n│  NEVER:                                                 │\n│    - Switch to unrelated task                           │\n│    - Retry >3 times without pivoting                    │\n│    - Abandon original task goal                         │\n└─────────────────────────────────────────────────────────┘\n```\n",{"name":5,"description":7},"main","6b84b0689ca0d9e7b1e4061188ff6b60758e0f8a","openspace\u002Flegacy\u002Fv1\u002Fgdpval_bench\u002Fskills\u002Fwrite-file-fallback-report-enhanced-4bc5c2",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":53,"prompt":63,"zipAvailable":64},[54,57,60],{"tool":55,"command":56},"npx","npx skills add HKUDS\u002FAgentPilot --skill context-anchored-fallback-report",{"tool":58,"command":59},"bunx","bunx skills add HKUDS\u002FAgentPilot --skill context-anchored-fallback-report",{"tool":61,"command":62},"pnpm","pnpm dlx skills add HKUDS\u002FAgentPilot --skill context-anchored-fallback-report","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fopenspace\u002Flegacy\u002Fv1\u002Fgdpval_bench\u002Fskills\u002Fwrite-file-fallback-report-enhanced-4bc5c2\u002FSKILL.md",true,[66],{"path":67,"resourceType":68,"resourceTypeDesc":69,"sizeBytes":70,"sourceHtmlUrl":71},".skill_id","OTHER","其他",56,"https:\u002F\u002Fgithub.com\u002FHKUDS\u002FAgentPilot\u002Fblob\u002Fmain\u002Fopenspace\u002Flegacy\u002Fv1\u002Fgdpval_bench\u002Fskills\u002Fwrite-file-fallback-report-enhanced-4bc5c2\u002F.skill_id",[73,92,111,135,158,179],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":39},"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",[],[75,80,81,37,82,83,84,85,86,87,88,89,90,91],"prepare","requests","missing","installation","rsi","up","use","user","work","SET","run","experiments",{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":39},"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",[],[94,99,100,101,102,103,104,105,106,107,108,109,110],"memory","review","auto","openspace","changes","facts","detect","entries","local","conflicts","stale","duplicates",{"skillNo":112,"name":113,"nameZh":6,"description":114,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":115,"categories":116,"keywords":122,"publishedAt":39},"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",[117],{"categoryNo":118,"categoryName":119,"taxonomyType":120,"taxonomyTypeDesc":121},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[123,124,125,102,126,127,128,129,130,131,132,133,134,32],"skill discovery","discovery","reusable","step by step","tokens","packages","help","cloud","reliability","capabilities","browse","search",{"skillNo":136,"name":137,"nameZh":6,"description":138,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":139,"categories":140,"keywords":142,"publishedAt":39},"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",[141],{"categoryNo":118,"categoryName":119,"taxonomyType":120,"taxonomyTypeDesc":121},[143,34,144,102,145,146,147,148,149,125,101,86,150,151,152,153,154,155,156,157],"delegate","delegate task","time","improve","desktop automation","worker","community","autonomous","tool","stack","for","share","tasks","automation","MCP",{"skillNo":159,"name":160,"nameZh":6,"description":161,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":162,"categories":163,"keywords":164,"publishedAt":39},"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",[],[165,33,166,167,168,169,170,171,30,172,173,174,175,176,153,177,178],"write special chars ts check","check","type","files","filters","only","template","python","typescript","output","Existing","type-checking","errors","script",{"skillNo":180,"name":181,"nameZh":6,"description":182,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":183,"categories":184,"keywords":185,"publishedAt":39},"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",[],[186,187,33,37,153,178,188,189,190,191,192,171,30,193,194,172],"write special chars readback","readback","helper","post","via","scripts","read","verification","cleanup"]