llms.txt for agents

Each tool has a clearly distinct role: llms_index discovers and parses the covering index for a URL, llms_search queries already-parsed indexes, and llms_page fetches individual docs pages. The descriptions explicitly guide when to use each, so there is no real ambiguity.

All names share the llms_ prefix and use lowercase snake_case, making the family obvious. However, llms_index and llms_page read as noun-like resources while llms_search is verb-like, so the names are not perfectly uniform in verb_noun structure.

Three tools is ideal for this server's scope: discover an index, search across indexes, and fetch a page. Each tool earns its place and the set does not feel thin or bloated.

The domain is llms.txt-aware documentation retrieval, and the tools cover the full workflow: find the relevant index, search it for the right page, and retrieve that page as markdown. llms_page can also fetch any URL, including llms-full.txt, so there are no obvious dead ends.