MCP server for Plannen — lets Claude Desktop and Claude Code interact with your local Plannen instance
Plannen MCP has 30 tools with consistent naming (verb_noun pattern) and complete input schemas. All tools have descriptions (10-200 chars, within baseline). However, output schemas are not documented in the source code, only input schemas are visible. Parameter descriptions are present but often generic (e.g., 'Event ID', 'Practice ID'). Error handling is minimal; no recovery guidance or actionable error messages visible. Tool composition is sound (each tool has one responsibility), but descriptions lack context about when to use each tool vs. similar ones (e.g., log_completion vs. log_activity distinction is buried in PLANNEN_INSTRUCTIONS, not in tool descriptions). No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels assigned in the spec.
Add items to an existing checklist
Mark a checklist item as checked/completed
Correct or update a previously recorded profile fact
Create a new checklist with optional initial items
Create a new event with title, dates, location, and optional enrollment details
Create a new memory/photo entry with optional caption and event association
Create a new practice/routine with frequency and description
Output schemas not documented. Tool descriptions state what is returned (e.g., 'Get a single event by ID') but do not specify the response structure, field names, or types. LLMs cannot plan downstream tool calls without knowing what fields to extract.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Risk levels are assigned in the spec (READ_ONLY, WRITE, DESTRUCTIVE, REVERSIBLE) but not exposed to the LLM via tool metadata. LLMs cannot infer safety constraints and may retry destructive operations.
Generic parameter descriptions. Many parameters use minimal descriptions ('Event ID', 'Practice ID', 'Fact ID') that do not explain format, constraints, or how to obtain the ID. Descriptions should include hints like 'UUID format' or 'Obtain from list_events() response'.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 60 | 2026-07-28+ | v2 |
Delete a checklist by ID
Delete an event by ID
Delete a memory/photo by ID
Delete a practice/routine by ID
Get a single checklist by ID with all items
Get a single event by ID with optional full details including sessions, parent, and memories
Get a single memory/photo by ID
Get a single practice/routine by ID
Get a single profile fact by ID
List checklists with optional filtering
List events with optional filtering by status, date range, group, and limit
List memories/photos with optional filtering by date range or event
List practices/routines with optional filtering
List profile facts about people, places, and preferences
Log an activity with optional duration or quantity measurement. Ticks matching active routine server-side
Log a completed task or activity. Resolves to: complete an existing open todo, mark a matching routine done, or log a new completed todo
Mark a checklist item as unchecked
Reverse the completion of a todo
Reverse marking a practice/routine as done
Update an existing event's properties
Update a memory/photo caption or metadata
Update an existing practice/routine
Create or update a durable fact about a person, place, or preference
No error recovery guidance. Tool descriptions do not explain what errors are possible, how to recover, or what to try next. E.g., create_event does not document what happens if start_date > end_date or if enrollment_deadline is in the past.
Ambiguous tool selection guidance. log_completion and log_activity both log user actions, but the distinction (duration/quantity vs. bare completion) is documented only in PLANNEN_INSTRUCTIONS, not in the tool descriptions themselves. LLMs may pick the wrong tool.