arborists

Each tool targets a distinct resource and action: get_* fetches one entity by slug, list_* enumerates a collection, and search_listings is the only text-based search. The two category tools (list_categories for companies vs list_product_categories for products) are clearly separated by descriptions. No overlapping purposes are present.

All tools follow a strict snake_case verb_noun pattern: get_course, get_listing, get_product, get_stats, list_*, search_listings. The verbs map cleanly to actions (get for single item, list for collections, search for filtered queries). There is no mixing of conventions.

11 tools is well within the ideal 3-15 range. Each tool corresponds to a distinct resource or operation (courses, listings, products, categories, places, sources, stats, search), with no redundant endpoints. The surface is appropriately scoped for a read-only data API.

The read-only surface covers the core domain: listings (get + search), courses (get + list), products (get + list + categories), places, categories, sources, and stats. Minor gaps exist—there is no dedicated list_listings (relies on search_listings), and no get_place/get_category/get_source tools—but agents can work around these via list endpoints.