Related Servers

Most tools map cleanly to a resource+action (worklog, plan, team, account, timesheet approval), and descriptions deliberately cross-reference siblings. The weak spot is the retrieval family: tempo_get_worklogs already filters by project/issue, yet separate tempo_get_worklogs_by_project/by_issue/by_user/by_team/by_account and tempo_search_worklogs exist, and get_accounts vs search_accounts / get_team_memberships vs search_team_memberships overlap similarly, so an agent can plausibly pick the wrong retriever.

Nearly every name follows tempo_<verb>_<noun> in snake_case with a predictable verb set (get, search, create, update, delete, submit, approve, reject, reopen, recall). The only non-verb name, tempo_healthcheck, is an obvious and isolated diagnostic outlier.

The Tempo domain is genuinely broad (worklogs, plans, teams, accounts, timesheet approvals, projects, periods, config), so a large surface is defensible. Still, 48 tools is heavy for a single MCP server, and roughly seven near-duplicate worklog retrieval tools plus several redundant search/get pairs inflate the set beyond what the workflows require.

Worklogs, plans, teams, accounts and the timesheet approval lifecycle have full or near-full CRUD coverage, which is strong. But team memberships are read-only (get/search with no add/update/remove), projects have only get/list, and account categories, roles, periods, schedules, work attributes and global config are read-only, leaving real gaps for administrative workflows.