Related Servers

Most tools are clearly distinguished by the agent_ or my_ prefix indicating the actor, and the operations are distinct (chats, contacts, messages, etc.). However, there is significant overlap between agent and user versions of similar operations (e.g., list_agent_chats vs list_my_chats, create_agent_chat vs create_my_chat), and the contact request handling is split differently for agents (unified respond) vs users (separate approve/reject/cancel), which could cause selection confusion in mixed contexts.

The dominant pattern is verb_noun with an agent_ or my_ prefix, but there are notable deviations: user contact request tools like create_contact_request, approve_contact_request, and reject_contact_request omit the my_ prefix, while resolve_handle and health_check are prefix-less. Additionally, messaging tools use different verbs (create_agent_chat_message vs send_my_chat_message), so the naming is not fully uniform.

At 44 tools, the surface is very large, largely due to duplicating nearly every operation for both agent and user contexts (e.g., ~20 agent tools and ~20 user tools). This feels like two separate sub-servers bundled together, inflating the count beyond what is necessary for a coherent set. The scope could be reduced by parameterizing the actor rather than creating parallel tools.

The tool set comprehensively covers chat management (create, get, list), participant management, messaging (send, list, events), contact management (add, remove, list, requests), message processing lifecycle (mark processing/processed/failed, list, get next), and profile/health checks for both agents and users. Minor gaps include lack of chat update/delete operations and no agent profile update, but these are not critical for the core workflows.