HeatCheck Agent API

The three tools map to distinct retrieval phases: orientation (about), discovery (search), and retrieval by id (fetch). There is minor overlap in that about also returns document ids and sample days, which search also surfaces, but the descriptions make the intended sequence clear enough to avoid misselection.

All names are single lowercase words with no casing or style mixing, which reads as a deliberate, predictable convention. The only wrinkle is that 'about' is a topic label while 'fetch' and 'search' are verbs, so it is not a strict verb_noun pattern.

Three tools is on the lean side but fits a read-only public data API: orient, query, and retrieve. No tool appears redundant, though a fourth (e.g. catalog/list by site) could plausibly earn its place.

For a read-only corpus the lifecycle is covered: discovery via search, id enumeration via about, and content retrieval plus free sample days via fetch. Gaps are minor, such as no explicit listing/browsing of all documents or site-scoped enumeration outside of search.