This server has moderate definition quality with consistent but minimal tool documentation. All 7 tools are properly registered with Zod schemas and basic descriptions. However, descriptions are terse (13-41 characters), parameter descriptions lack depth, and output schemas are not formally documented. No tool annotations (readOnlyHint, destructiveHint) are present despite clear WRITE and DESTRUCTIVE operations. Error handling is present but generic. The server follows basic naming conventions (verb_noun) but lacks the rich context production tools need for reliable LLM invocation.
Tool to create a new calendar
Tool to create a new event in a calendar
Tool to delete a calendar
Tool to delete an event
Tool to get events from a specific calendar
Tool to get all calendars
Tool to update an existing event
Tool descriptions are under 50 characters and lack actionable context. Examples: 'Tool to get all calendars' (27 chars), 'Tool to delete an event' (23 chars). LLMs cannot distinguish when to use these tools vs alternatives or understand consequences without richer context.
Parameter descriptions are missing or generic. The 'color' parameter in createCalendar has no description (only marked 'optional' in tool description). LLMs cannot infer valid color formats (hex, names, enums?) without explicit guidance.
Output schemas are completely undocumented. Tools return JSON wrapped in text content (e.g., '{calendars: [...]}' or '{success: true, message: ..., event: ...}') but the structure is never formally specified. LLMs cannot plan downstream operations or extract fields reliably.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 56 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 0 | - | v1 |
Destructive tools (deleteCalendarEvent, deleteCalendar) lack confirmation or dry-run support. Agents cannot preview consequences before triggering irreversible deletions. No idempotent or destructive hints are declared.
Tool annotations (readOnlyHint, destructiveHint, idempotentHint) are completely absent. The MCP spec supports these to guide agent planning, but this server declares none. Agents cannot reliably distinguish safe reads from mutations.
Date format validation is incomplete. The createCalendarEvent and updateCalendarEvent tools describe dates as 'ISO8601 format' but do not validate or guide the LLM on the exact expected format. Error messages reference unmet requirements but don't specify the correct format.
Error handling is generic and does not guide recovery. Example: 'Failed to create event' provides no actionable next step. The code has special-case date error handling for createCalendarEvent/updateCalendarEvent but the error messages still don't specify the correct ISO8601 variant required.