Search and navigate external documentation sites via llms.txt and MCP endpoints.
npx skills add https://github.com/mintlify/docs --skill doc-reader
Please help me install this Agent Skill. Skill: doc-reader Source: https://github.com/mintlify/docs/tree/main/.claude/skills/doc-reader Command: npx skills add https://github.com/mintlify/docs --skill doc-reader
This skill helps users efficiently consume and navigate external documentation sites without overwhelming their context or missing important details.
To use this skill, begin by visiting the documentation site and:
High-intent search queries and answers about installing and using this skill.
You can search external documentation efficiently by leveraging llms.txt files and optional MCP endpoints to perform semantic search, keyword grep, or chunked loading of relevant sections. This approach prevents context overload by loading only the specific sections you need.
The best way to find exact terms is to grep the llms-full.txt file if the documentation site provides it at the root. This allows you to perform precise keyword matching across the entire documentation content quickly.
To use MCP endpoints for semantic search, check the documentation site root for an MCP configuration, connect using the documented settings, and then execute targeted queries. This enables fast discovery of relevant APIs, libraries, or concepts.
Yes, this approach is adaptable to multiple docs sites, enabling quick orientation to various APIs, libraries, or tools. You can view the HTML structure to understand layout and navigation, then apply your chosen consumption strategy across different sources.
Use MCP semantic search for targeted questions about specific APIs or concepts, and use chunked loading or grep on llms-full.txt when you need to explore broader sections or find exact terms. Choose your strategy based on whether you need conceptual answers or exact matches.
fabricioctellesPublish an LLM-friendly website overview at the standard llms.txt endpoint.
ctxfstTransform Markdown documents into CtxFST-formatted semantic chunks with frontmatter metadata.
pradyumnasagarGuide authors through creating and validating ORS SKILL.md files.
Cogni-AI-OUFetch llms.txt and markdown fallbacks to standardize website documentation context.
