No user-submitted related servers found.
Scored across 5 tools
Each tool has a clearly distinct lifecycle role: model catalog, starting a new conversation, replying to an existing one, reading/polling output, and cancelling a specific turn. The start-vs-reply boundary is well clarified by 'independent' new conversation versus 'existing discussion', so misselection is unlikely.
All names are snake_case and follow a predictable resource_action pattern: models_list, discussion_start, discussion_reply, discussion_read, discussion_cancel. The only minor variation is plural 'models' versus singular 'discussion', which is natural and still consistent.
Five tools map directly to the necessary catalog lookup plus discussion lifecycle operations. There is no obvious bloat, and the surface is neither too thin nor too heavy for this purpose.
The core lifecycle is covered: model discovery, start, reply, read/poll, and cancel. Minor gaps exist, such as no tool to list or discover existing discussions and no explicit finish/archive/delete operation, but agents can work around these when they track discussion IDs and read turn status.