跳到正文
cccskills
免费在 GitHub 公开

search-fallback-workflow

Continue suitable tasks when web search repeatedly fails by assessing whether existing knowledge is sufficient. Mark unverified claims, distinguish confidence levels, and provide a follow-up checklist so readers can see what still needs checking.

中文简介正在准备中。当前显示英语简介。

原文查看原文说明

Graceful degradation workflow for continuing tasks when web search tools fail, using internal knowledge with verification caveats

查看安装方法

适用场景

  • Assessing whether work can continue without web search
  • Drafting documents from established knowledge during search outages
  • Marking unverified claims and confidence levels
  • Preparing a checklist for later source verification

包含的文件(2)

  • SKILL.md4.6 KB
  • .skill_id38 B

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Search Fallback Workflow

When to Apply

Use this workflow when:

  • search_web tool fails repeatedly (typically 2+ consecutive failures)
  • External information is needed but unavailable due to tool limitations
  • Task completion is still possible with reasonable confidence using internal knowledge
  • The output can include appropriate caveats about verification needs

Core Pattern

  1. Detect Search Failure

    • Monitor search_web responses for errors, empty results, or timeouts
    • After 2 consecutive failures, assess whether the task can proceed without fresh data
  2. Assess Feasibility

    • Determine if internal knowledge is sufficient for the task at hand
    • Identify which sections/claims require external verification
    • Evaluate risk level: Can the task proceed safely with caveats?
  3. Proceed with Internal Knowledge

    • Continue the task using well-established facts, regulations, or principles
    • Prioritize accuracy over completeness when uncertain
    • Use conservative language for claims that cannot be verified
  4. Flag for Verification

    • Clearly mark any claims that should be verified:
      [VERIFICATION NEEDED: Search tool unavailable during generation]
      
    • Include a verification checklist in the execution summary
    • Distinguish between high-confidence and low-confidence assertions
  5. Document Limitations

    • Include an explicit note in the output about search unavailability
    • List specific items requiring follow-up verification
    • Provide recommendations for manual verification steps

Implementation Template

## Execution Notes

**Search Tool Status**: Unavailable during generation

**Confidence Levels**:
- ✓ High confidence: Established regulations, well-documented facts
- ⚠ Medium confidence: Industry standards, commonly accepted practices  
- ✗ Requires verification: Time-sensitive data, recent changes, specific citations

**Verification Checklist**:
- [ ] Verify citation: [specific reference]
- [ ] Confirm current status: [specific item]
- [ ] Review recent updates: [specific topic]

Code Pattern for Agents

# Pseudocode for search fallback detection
search_attempts = 0
max_search_attempts = 2
search_failed = False

for query in required_searches:
    result = search_web(query)
    if result.success:
        process_result(result)
    else:
        search_attempts += 1
        if search_attempts >= max_search_attempts:
            search_failed = True
            log_warning("Search tool unavailable, proceeding with fallback")
            break

if search_failed:
    # Apply graceful degradation
    content = generate_with_internal_knowledge()
    add_verification_flags(content)
    document_limitations_in_summary()

Example Output Structure

When using this fallback, structure documents with clear sections:

# Document Title

## Disclaimer
This document was generated with limited access to external verification tools.
Items marked with [VERIFICATION NEEDED] should be confirmed before final use.

## Content Sections
[Standard content with internal knowledge]

## Items Requiring Verification
- Topic A: Current regulatory status needs confirmation
- Topic B: Recent policy changes should be checked
- Topic C: Specific citations require source validation

## Recommendations
1. Cross-reference with official sources
2. Verify time-sensitive information
3. Consult subject matter experts for critical decisions

Best Practices

  1. Don't fabricate citations - Better to flag as unverified than invent sources
  2. Use conservative language - "Generally," "Typically," "Commonly" instead of absolutes
  3. Prioritize safety - If misinformation could cause harm, recommend manual verification
  4. Track what's missing - Maintain a clear list of items needing follow-up
  5. Be transparent - Clearly communicate limitations to end users

When NOT to Use This Pattern

  • Tasks requiring current/real-time data (stock prices, news, weather)
  • Legal or medical advice requiring authoritative sources
  • Compliance documentation where citations are mandatory
  • Situations where incorrect information could cause significant harm

Related Patterns

  • citation-management: Handling references and sources in documents
  • error-recovery-workflow: General patterns for tool failure recovery
  • confidence-annotation: Marking certainty levels in generated content

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Prepare audio stems, the separate parts of a track, for mixing by matching their durations with silence, loops, or crossfades. Verify each generated and processed file before moving to the next production step.

中文简介正在准备中。当前显示英语简介。

  • Match audio parts with different durations
  • Extend drums and melodic parts using suitable methods
  • Check audio specifications before mixing
HKUDS/OpenSpace77622026年8月13日 更新

Produce audio in separate stems, meaning individual instrument tracks, and verify each output as you go. Align track lengths with silence, loops, or crossfades, then apply effects, mix a master, and package the files.

中文简介正在准备中。当前显示英语简介。

  • Align instrument tracks to a reference audio file
  • Choose how to extend tracks that are too short
  • Verify audio properties after generation and effects
HKUDS/OpenSpace77622026年8月13日 更新

Build serverless endpoints that relay external API requests for a dashboard while keeping API keys on the server. Includes Vercel examples for input checks, caching, cross-origin access, and consistent JSON responses.

中文简介正在准备中。当前显示英语简介。

  • Connect a dashboard to external APIs without exposing API keys
  • Create endpoints for stock quotes and news
  • Standardize API responses and error handling
HKUDS/OpenSpace77622026年8月13日 更新

Produce a master track and separate audio parts through a structured Python workflow. Apply filters and reverb, export WAV files with consistent settings, package them in a ZIP archive, and check that the deliverables meet the requested specifications.

中文简介正在准备中。当前显示英语简介。

  • Export a master track with separate audio parts
  • Apply filters and reverb to audio
  • Check duration and format across deliverables
HKUDS/OpenSpace77622026年8月13日 更新

Keep document creation moving when web searches and page retrieval repeatedly fail. The skill preserves the original task, uses the model’s existing knowledge, and clearly identifies information that could not be checked against external sources.

中文简介正在准备中。当前显示英语简介。

  • Continue drafting after repeated web retrieval failures
  • Preserve task requirements during network errors
  • Mark information that could not be externally verified
HKUDS/OpenSpace77622026年8月13日 更新

When Python execution repeatedly fails in execute_code_sandbox, use a file-based alternative: save the script, run it through the shell, and capture results or errors. The guidance covers paths, imports, timeouts, structured output, and optional cleanup.

中文简介正在准备中。当前显示英语简介。

  • Retry Python work through the shell after repeated sandbox failures
  • Run longer tasks with an adjusted timeout
  • Execute Python work that needs multiple files or system utilities
HKUDS/OpenSpace77622026年8月13日 更新

查看 HKUDS 的全部技能

报告此技能的问题