NewsMAN MCP by usefulapi

Most tools are cleanly separated by resource + action (lists, segments, subscribers, newsletters, automations, imports), and WRITE tools are explicitly flagged. Two mild overlaps remain: list_subscribers vs search_subscribers (search with no keywords duplicates the list call) and get_subscriber (by email) vs get_subscriber_by_id, though the descriptions do disambiguate them.

Nearly every tool follows a newsman_verb_noun pattern (get_list, list_newsletters, add_subscriber_tags, unsubscribe_email), which is easy to scan and predictable. Minor deviations: the noun varies between 'email' and 'subscriber' for the same entity (subscribe_email / update_subscriber), and one tool carries a 'by_id' qualifier while its sibling does not.

At 26 tools the surface is on the heavy side of the guidance, but the domain genuinely spans six resources (lists, segments, subscribers, newsletters, automations, imports), so most tools map to a distinct endpoint. It is still more than an agent can comfortably hold at once and a few count/stats tools could be folded into their list counterparts.

Subscriber lifecycle is strong (subscribe, update, unsubscribe, tag, segment membership, history), but the surface is read-only for lists, segments, newsletters and automations, with no create/update/delete for those resources and no way to remove tags. get_import_status is a dead end since no tool starts an import, leaving an obvious gap.