Tool names are mostly distinct, but several overlap in purpose: participant_stats vs. meeting_stats, search_meetings vs. search_content, and get_meeting vs. get_transcript vs. export_meeting. Without any descriptions, an agent must guess the exact boundaries, making misselection likely in these clusters.

All tool names use consistent snake_case and follow a clear verb_noun or noun_noun pattern: list_meetings, get_meeting, export_meeting, search_meetings, create_webhook, delete_webhook, etc. The two stats tools (participant_stats, meeting_stats) are noun phrases but still fit the predictable naming style.

Fourteen tools is well within the ideal range for a meeting-intelligence server. The set covers core read, search, export, stats, team, and webhook operations without obvious bloat or missing essentials.

The surface covers meetings, transcripts, exports, search, stats, action items, teams, and webhooks, which is strong for a read-oriented recording platform. Minor gaps exist in webhook management (no list/get/update) and deeper team or participant detail, but agents can work around them.