Loading ToolBench
TickTick Model Context Protocol Server for Claude
The TickTick MCP server has 21 well-named tools with consistent verb_noun naming (list_, get_, create_, update_, delete_, complete_, search_). All tools have descriptions and input schemas. However, quality is uneven: naming is strong; descriptions are adequate but often generic; parameter descriptions vary widely; output schemas are completely undocumented; error handling is not visible in the provided code. No destructive operation has a confirmation mechanism. Per-tool scores range from 55 (minimal schema) to 75 (rich schema like create_task). Average across all 21 tools is 62, placing this in the fair-to-good range for community servers.
Archive (close) a project
Check TickTick authentication status
Start TickTick OAuth authorization flow
Mark a task as completed
Create a new project
Create a new task
Delete a project
No output schemas documented. Tool descriptions state what they return (e.g. 'List all projects', 'Get project statistics') but the response structure is completely undocumented. LLMs cannot plan downstream tool calls or extract specific fields without knowing what data is returned.
Destructive operations (delete_project, delete_task) lack any confirmation mechanism or dry-run option. Agents can execute irreversible deletions without a chance to verify intent. There is no pattern:confirmation-request implementation.
Five tools with empty input schemas (no properties defined): ticktick_auth_status, ticktick_authorize, ticktick_logout, ticktick_health_check. These tools accept no parameters but the schema is minimal. While intentional, LLMs benefit from richer schemas explaining what the tool does.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | C | 66 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 35 | - | v1 |
Delete a task
Get a specific project by ID
Get project statistics (task counts, completion rates, etc.)
Get all tasks in a specific project
Get a specific task by ID
Check TickTick API connection health
List all projects with optional filtering
List all tasks with optional filtering and sorting
Logout from TickTick and clear stored tokens
Search tasks with comprehensive filtering and sorting options
Unarchive (reopen) a project
Mark a task as not completed
Update an existing project
Update an existing task
No pagination documented. Tools like ticktick_list_projects, ticktick_list_tasks return potentially large lists but lack limit, offset, or cursor parameters. Without pagination, large result sets blow context windows. Descriptions do not mention result limits or pagination strategy.
No error handling examples or recovery guidance visible. If a tool fails (e.g. invalid project ID, auth failure), the error response structure and how an agent should respond is not documented.
Tool descriptions are generic and often under-optimized for LLM selection. E.g., 'Check TickTick authentication status' is vague, does it return a boolean, a user object, or a token status? 'Get project statistics', which statistics? Completion rate? Task count? Time tracking?