No user-submitted related servers found.
Scored across 4 tools
Each tool targets a distinct step: discover collections, query a panel, submit a judgement, and enumerate unjudged seats. The only soft boundary is between `verdict` and `pending_verdicts`, which the description partially resolves by telling the agent it never needs to call the latter, though that still invites hesitation.
`list_collections` uses a verb_noun pattern while `ask` and `verdict` are bare action words and `pending_verdicts` is a noun phrase. The names are readable and memorable, but there is no single predictable convention an agent can rely on.
Four tools is lean but matches the narrow discover→ask→judge loop, with no redundant entries. `pending_verdicts` exists mainly for a plugin hook rather than agent use, so the effective agent-facing surface is even smaller, which is acceptable but slightly thin.
The core lifecycle is covered: enumerate collections, fan a prompt to all seats, and record per-seat correctness feedback that feeds routing. Gaps are minor: no way to inspect prior calls or verdicts, cancel a long-running `ask`, or query the 'confidently wrong' record that verdicts supposedly populate.