No user-submitted related servers found.
Scored across 26 tools
The set has several intentionally overlapping pairs (context.get vs context.read, session.record vs event.append/decision.create/context.update, session.checkpoint vs session.record, memory.put vs memory.promote), but each description explicitly states when to prefer which, which keeps misselection low. A few boundaries (e.g. context.update task fields vs task.update) still require reading two descriptions to resolve, so it is not a clean 5.
Every tool follows a strict dotted `resource.action` convention (artifact.get, decision.list, blocker.resolve, context.summarize, memory.search, session.checkpoint, task.update, workspace.diff). Namespaces are used consistently and verbs are predictable across all 26 tools.
26 tools is heavy for a session/context-tracking server, and several exist mainly as thin duplicates of others (context.get/context.read, context.update vs task.update, memory.promote vs memory.put, session.record vs the individual record tools). The surface is defensible but feels broader than the domain warrants.
Core lifecycle is well covered (events, decisions, blockers, tasks, checkpoints, memory, workspace), but there are notable gaps: no event.list/query despite an append-only log (events are only visible via context.summarize's last 8), no task.list, and no way to update or delete a decision or edit a blocker besides resolve. Agents will hit dead ends when they need to inspect history or correct prior entries.
