doc-research

Original🇺🇸 English
Translated

This skill should be used when the user asks to "search documentation", "find docs for", "look up API reference", "how to use [library]", or needs up-to-date information about a library, framework, or technology. Performs prioritized documentation research using c7 skill, cloudflare-docs, WebSearch, and cloudflare-browser.

2installs
Added on

NPX Install

npx skill4agent add totto2727-dotfiles/agents doc-research

Tags

Translated version includes tags in frontmatter

Documentation Research

Perform prioritized documentation research and return only relevant results.

Available Tools

PriorityToolUse Case
Special
mcp__plugin_totto2727_cloudflare-docs__search_cloudflare_documentation
Cloudflare product queries only
Primary
c7
skill (via Skill tool)
General library/framework official docs
Fallback
WebSearch
When above sources are insufficient
Fallback
mcp__plugin_totto2727_cloudflare-browser__get_url_markdown
Extract clean markdown from URLs found via WebSearch

Workflow

  1. Determine query type
    • Cloudflare-related → Use
      search_cloudflare_documentation
    • General library/framework → Invoke
      c7
      skill via Skill tool
  2. Evaluate results
    • Sufficient information found → Return results
    • Insufficient → Proceed to step 3
  3. Fallback: WebSearch
    • Search with keywords targeting official sources (official docs, official blogs)
    • Prefer results from official documentation sites
  4. Deep content extraction (if needed)
    • Use
      get_url_markdown
      to extract clean markdown from promising URLs
    • Only fetch URLs that are likely to contain the needed information

Guidelines

  • Always start with the highest-priority source for the query type
  • Prefer official documentation over third-party content
  • Return concise, relevant results only — do not include excessive raw output
  • When using WebSearch, construct queries that target official sources
  • When multiple results are found, summarize the key information rather than dumping raw content