Agent Skills: Context7 Skill

Access up-to-date, version-specific documentation and code examples from Context7. Use this skill to verify library and framework details.

UncategorizedID: ceshine/ceshine-agent-skills/context7-skill

Install this agent skill to your local

pnpm dlx add-skill https://github.com/ceshine/ceshine-agent-skills/tree/HEAD/context7-skill

Skill Files

Browse the full folder contents for context7-skill.

Download Skill

Loading file tree…

context7-skill/SKILL.md

Skill Metadata

Name
context7-skill
Description
Access up-to-date, version-specific documentation and code examples from Context7. Use this skill to verify library and framework details.

Context7 Skill

Overview

The Context7 skill connects you to accurate, version-specific documentation and code examples directly from the source. Use this skill to verify syntax, API details, and usage patterns for third-party libraries and frameworks, ensuring your code is based on the correct version.

Access Methods:

  1. MCP Tools (Primary): Direct calls to resolve_library_id and query_docs.
  2. CLI Fallback (Secondary): A Python script (scripts/context7_cli.py) using the FastMCP v2 client, intended for use when direct MCP tools are unavailable.

Prerequisites

  • Context7 API Key: Must be set as the environment variable CONTEXT7_API_KEY.
  • Python Manager (uv): Required only if using the CLI fallback script.

Available Tools

1. Library Resolver (resolve-library-id)

Resolves a package or library name to a Context7-compatible library ID and returns a list of matches.

  • MCP Call: resolve_library_id(query="...", libraryName="...")
  • CLI Command: uv run --script scripts/context7_cli.py resolve-library-id <name>

2. Documentation Query (query-docs)

Retrieves documentation and code examples using a specific library ID.

  • MCP Call: query_docs(libraryId="/org/project/version", query="...")
  • CLI Command: uv run --script scripts/context7_cli.py query-docs <library_id> <query>

IMPORTANT: Tool names may have prefixes (e.g., context7_resolve_library_id) depending on the runtime environment. Always check available tools first.

Usage Guidelines

  1. Resolve First: Always obtain a valid library ID via resolve_library_id before querying, unless the user provides a full ID (e.g., /org/project/version).
  2. Limit Attempts: Do not retry a tool call more than three times for the same query. If unsuccessful, proceed with the best available information.

Workflow

Step 1: Check Availability

Determine if the resolve_library_id and query_docs tools are directly available in your environment. If not, default to the CLI fallback commands.

Step 2: Resolve Library ID

Use resolve-library-id to identify the correct library.

Selection Criteria:

  • Exact Match: Prioritize names that exactly match the user's request.
  • Relevance: Ensure the description aligns with the user's intent.
  • Quality: Look for high documentation coverage (snippet counts), reputation, and benchmark scores.

Action:

  • If ambiguous, ask the user for clarification.
  • Briefly explain the selected library to the user.
  • If no good match is found, clearly state this and suggest query refinements.

Step 3: Query Documentation

Use query-docs with the resolved libraryId.

Handling Results:

  • The tool returns a snippet or summary.
  • Example Output:
    Source: https://github.com/context7/react_dev/blob/main/learn.md
    ... (content) ...
    
  • Insufficient Info? If the returned text is incomplete, use a web fetch tool (if available) to retrieve the full content from the provided source URL.

Configuration

| Variable | Description | Required | | :----------------- | :---------------------------------------- | :------- | | CONTEXT7_API_KEY | API key for authenticating with Context7. | Yes |

Resources

  • scripts/context7_cli.py: Unified CLI entry point for fallback access.
  • references/troubleshooting.md: Solutions for common integration issues.