MCPSkillHub

Each tool maps to a clearly distinct action on the same resource: search (find many), get (retrieve one by slug), compare (diff 2-5). The descriptions reinforce the boundaries, including input constraints (slug vs keyword vs 2-5 servers), so an agent can pick correctly without guessing.

All three follow a strict verb_noun snake_case pattern (search_servers, get_server, compare_servers) with a consistent 'server(s)' resource noun. The singular/plural variation tracks single-resource vs multi-resource operations, which is a predictable convention rather than an inconsistency.

Three tools is on the lean side but each covers a genuinely different discovery need (search, detail, comparison), so nothing feels redundant. A browse/list-all or category-listing tool could reasonably earn a fourth slot, keeping it just below a fully rounded set.

The core registry workflow — find by keyword, inspect install/env/score details, then compare candidates — is fully covered with no dead ends. Minor gaps exist: no way to enumerate or browse the catalog without a keyword, and no category/tag filtering, which an agent would have to work around via broad searches.