| name | research-docs |
| description | Pull live, up-to-date library and API documentation. Use when: asking how to use any external package, framework, or API; before implementing against a third-party SDK; when training data may be stale (anything released after 2023); looking up config options, migration guides, or changelogs. Uses context7 MCP. |
| argument-hint | Library name and what you need to know (e.g. 'drizzle-orm query builder') |
Research Docs via context7
Training data goes stale. context7 pulls live documentation. Always use it before answering questions about external libraries — never rely on memory alone.
When to Use
- "How do I configure X in library Y?"
- "What's the correct API for Z in version N?"
- Anything involving a package that may have changed since 2023
- Migration guides (v1 → v2, etc.)
- Before implementing a new integration
Tools
mcp_wulfnet-conte2_context7_get_tool_schema(tool_name) # schema lookup
mcp_wulfnet-conte2_context7_invoke_tool(tool_name, tool_input)
Available tools (from schema): resolve-library-id, query-docs
Procedure
Step 1 — Resolve the library ID
tool_name: "resolve-library-id"
tool_input: {
"query": "<plain-english library name or description>",
"libraryName": "<npm/pypi/etc package name if known>"
}
Returns a libraryId. Use the exact value in step 2.
Step 2 — Pull the docs
tool_name: "query-docs"
tool_input: {
"libraryId": "<id from step 1>",
"query": "<specific thing you need to know>"
}
Step 3 — Apply and cite
- Quote or summarize the relevant section in your response
- If the docs contradict your training data, trust the docs
- If the docs are incomplete, note that and fall back to
context-mode for web research
Rules
- Always run this before answering any "how do I use library X" question
- Never invent API signatures from memory for libraries you haven't confirmed
- If
resolve-library-id returns nothing useful, try web-task skill instead
