AgenticJobs

Each tool has a stated distinct role: search_jobs for filtered queries, vibe_top for unfiltered top-N, fetch_jobs for batch hydration by id, get_job for a single full record with scoring rationale, guide and stats as metadata. The only real friction is search_jobs vs vibe_top (the latter is essentially a prefiltered search) and fetch_jobs vs get_job, plus conflicting 'call this first' guidance from both guide and search_jobs.

Three tools follow a clean verb_noun pattern (fetch_jobs, get_job, search_jobs), but guide, stats, and vibe_top are bare nouns or noun phrases with no verb. Still readable and predictable enough that an agent won't misread them, but the convention is mixed.

Six tools is well within the sweet spot and each one earns its place: discovery (search_jobs, vibe_top), hydration (fetch_jobs, get_job), and orientation (guide, stats). No filler tools and no need to split or merge anything.

The core lifecycle of a job-index server is covered: orient, search, rank, batch-fetch, detail, and aggregate stats. Gaps are minor — there's no explicit pagination/cursor beyond `limit`, no tool to enumerate available sources or filterable fields (only stats counts and prose in guide), which agents could work around but would benefit from.