Awesome-omni-skill context7
Library documentation lookup with tokenization and reranking. Use when user mentions external libraries, frameworks, or packages (e.g., how do I use React hooks, FastAPI authentication, pandas DataFrame). Triggers on library names, package imports, framework questions. Returns top 5 most relevant documentation chunks.
install
source · Clone the upstream repo
git clone https://github.com/diegosouzapw/awesome-omni-skill
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/diegosouzapw/awesome-omni-skill "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/development/context7-zach-source" ~/.claude/skills/diegosouzapw-awesome-omni-skill-context7-c33a8c && rm -rf "$T"
manifest:
skills/development/context7-zach-source/SKILL.mdsource content
Context7 - Library Documentation
Fetch up-to-date documentation for any library with intelligent chunking and reranking.
Workflow
Step 1: Resolve Library ID
First, resolve the library name to a Context7-compatible ID:
mcp__context7__resolve-library-id({ libraryName: "<package name>" })
Example:
libraryName: "react" → returns /facebook/react
Step 2: Get Documentation
Fetch docs with topic focus and token budget:
mcp__context7__get-library-docs({ context7CompatibleLibraryID: "<library_id>", topic: "<specific topic>", tokens: 5000 })
Step 3: Rerank Results
After receiving documentation, apply this mental reranking:
- Score by relevance - Prioritize chunks containing query terms
- Consider recency - Newer API patterns over deprecated ones
- Focus on examples - Code snippets > prose descriptions
- Limit to top 5 - Present only the 5 most relevant chunks
Quick Reference
| Library Type | Example Query | Topic |
|---|---|---|
| React/Vue/Angular | "React hooks for state" | "hooks" or "useState" |
| FastAPI/Flask/Django | "FastAPI authentication" | "auth" or "security" |
| Pandas/NumPy | "pandas merge dataframes" | "merge" or "join" |
| SQLAlchemy/Prisma | "SQLAlchemy relationships" | "relationships" |
| Testing libs | "pytest fixtures" | "fixtures" |
Token Budget Guidelines
- Quick lookup: 2000-3000 tokens
- Deep dive: 5000-8000 tokens
- Comprehensive: 10000+ tokens
Reranking Script
For programmatic reranking, use the wrapper script:
# Process context7 output and rerank python ~/.config/context7/context7_wrapper.py process \ --query "authentication middleware" \ --top 5 \ --input docs.md
Error Handling
If library not found:
- Try alternative names (e.g., "nextjs" vs "next.js")
- Check for organization prefix (e.g., "/vercel/next.js")
- Search for related libraries
Examples
React Hooks
1. resolve-library-id({ libraryName: "react" }) 2. get-library-docs({ context7CompatibleLibraryID: "/facebook/react", topic: "hooks", tokens: 5000 })
FastAPI Auth
1. resolve-library-id({ libraryName: "fastapi" }) 2. get-library-docs({ context7CompatibleLibraryID: "/tiangolo/fastapi", topic: "security oauth2", tokens: 5000 })