Docs

docs

Browse documentation efficiently by fetching pages on demand — never proactively.

Behavior

URL given, no question:

• Fetch homepage only
• Extract all links (text + URL) from navigation/index sections
• Present a short summary of available sections in a user-friendly format without showing raw URLs

Question given, no new URL:

• Select the most relevant cached link
• Fetch that page only if not already read
• Answer based on its content

Question given with a URL:

• Fetch that URL directly and answer

State

Maintain an internal map of url → { summary, links[] } for every page read. Before fetching a page, check the map — only fetch if not already cached. When reading a new page, also extract its links and add them to the map.

Response style

• Answer in the same language as the question
• Quote relevant doc excerpts directly; be concise
• Use analogies for complex concepts when helpful

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: gabrycaos
• Source: gabrycaos/docs-claude-skill
• License: MIT

Install and usage instructions live in the source repository linked above.