No user-submitted related servers found.
Scored across 45 tools
Most tools target a clearly distinct resource+action (record_claim vs supersede_claim vs retract_claim, watch/unwatch/list_watches, next/complete pairs for jobs), and descriptions explicitly disambiguate near-neighbours like timeline vs what_changed vs recent_records. A few pairs still blur: whatif vs impact are both dry-run reaches, and brief vs recall vs digest all surface prior context, though the descriptions do differentiate them on close reading.
There is a solid verb_noun family (record_claim, record_thesis, update_thesis, merge_entities, resolve_alert, complete_mapping) and consistent next_/complete_ job pairs, but it sits alongside many bare single-word tools (brief, recall, search, neighbors, timeline, impact, audit, digest) and noun-phrase tools (alert_stats, thesis_weak_points, filing_diff). The mixed style is readable and roughly maps verbs to mutations and nouns to read-only views, but it is not a single predictable convention.
45 tools is far above the ergonomic range and puts heavy pressure on tool selection even with distinct descriptions. The domain is broad (graph research, alerting, job leasing, filings, memos, scorecard), so each tool is plausibly motivated, but the surface would benefit from consolidation or grouping.
Coverage is unusually thorough: full entity/claim/thesis/relation/evidence lifecycle, corrections with history, event-driven alerts, watchlist management, job leasing for distillation and mapping, memos, explorer export, scorecard and audit. Minor gaps exist — there is no direct get_thesis or list_theses retrieval tool, with thesis inspection routed through brief, recall, thesis_memo and update_thesis.