smithery/neversight

doc-search

Automatically activates when documentation lookup is needed. Routes to the optimal source (Sanity MCP, Context7, Ref, or WebSearch) based on the technology being queried.

Installation

$ npx skills add smithery/neversight --skill doc-search

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from smithery/neversight · top by installs.

npx skills add smithery/neversight

Browse all from smithery/neversight

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 3,754 B
  • docs SUMMARY.md 188 B

History

  1. First recorded snapshot · 0 installs

SKILL.md

Smart Documentation Search

Purpose

This skill automatically activates when you need to look up documentation, API references, or best practices. It intelligently routes to the best available source.

When to Activate

Invoke this skill when the request involves:

  • Looking up how to use a library/framework feature
  • Finding API references or function signatures
  • Searching for code examples or patterns
  • Checking best practices or conventions
  • Verifying correct usage before writing code

Routing Logic (Priority Order)

1. Native MCP Tools (if available)

Always check if a dedicated MCP server exists for the technology. MCP tools provide the most accurate, integrated results.

Technology MCP Tools Use For
Sanity mcpSanitysearchdocs, mcpSanityreaddocs, mcpSanitygetgroqspecification, mcpSanitygetsanityrules Schemas, GROQ, Studio, Visual Editing
GitHub mcpgithub* (if configured) Repos, issues, PRs
Notion mcpnotion* (if configured) Workspace docs
Other MCPs Check available tools Varies

How to detect: If mcp[Technology] tools are available, use them first.

2. Context7 (Popular Libraries)

Use for well-indexed popular libraries. Context7 provides curated, high-quality code snippets.

Step 1: mcp__context7__resolve-library-id (get library ID)
Step 2: mcp__context7__get-library-docs (fetch docs)

Pre-resolved Library IDs (skip Step 1 for these):

Technology Context7 ID
React Router /remix-run/react-router
React /websites/react_dev
Next.js /vercel/next.js
Tailwind CSS /tailwindlabs/tailwindcss
TypeScript /microsoft/typescript
Prisma /prisma/prisma
Zod /colinhacks/zod
Vite /vitejs/vite
Astro /withastro/astro
tRPC /trpc/trpc
Drizzle /drizzle-team/drizzle-orm
shadcn/ui /shadcn-ui/ui

Parameters:

  • topic: Focus the search (e.g., "loaders", "hooks", "validation")
  • mode: "code" for API/examples, "info" for concepts/architecture

3. Ref (Broader Search)

Use for broader searches across multiple sources.

Step 1: mcp__Ref__ref_search_documentation (search)
Step 2: mcp__Ref__ref_read_url (read specific URLs)

When to use:

  • Library not found in Context7
  • Need to search private repositories (add ref_src=private)
  • Looking for blog posts, tutorials, or discussions
  • Obscure or niche libraries

4. WebSearch (Fallback)

Use WebSearch as last resort for:

  • Very recent information (current year)
  • Comparing multiple solutions
  • Community discussions or Stack Overflow answers

Decision Flowchart

Is there an MCP for this technology?
├─ YES → Use MCP tools
└─ NO → Is it a popular library?
         ├─ YES → Use Context7
         └─ NO/Not found → Use Ref
                           └─ Still not enough? → WebSearch

Response Format

After retrieving documentation:

  1. Source: Which tool was used and why
  2. Summary: Concise answer to the query
  3. Code Examples: Relevant snippets (if applicable)
  4. Links: Source URLs for further reading

Important Rules

  • NEVER guess - always verify with documentation
  • Check MCP availability first - most accurate source
  • Use topic parameter in Context7 to focus results
  • Combine sources when query spans multiple technologies
  • Cite sources with URLs in your response