Loading ToolBench
MCP server for accessing Wrike DataHub databases, spaces, fields, and records
Server has 7 tools with mostly complete input/output schemas and descriptions. Naming follows verb_noun convention (list_, get_, create_). However, critical gaps exist: (1) parameter descriptions are sparse or missing in several tools (e.g., listSpacesTool has empty inputSchema with no properties); (2) output schemas have type errors (getDatabaseTool uses string literals instead of {type:"string"}); (3) descriptions lack actionable context for LLM selection (e.g., 'List spaces' vs 'List spaces to discover available workspaces'); (4) no error handling guidance or recovery paths documented; (5) create_database tool has deeply nested config schema but lacks clear documentation of field type dependencies. Pagination is present (nextPageToken) but not consistently documented across all list tools.
Create a new database with specified fields
Get detailed information about a specific database efficiently including record count
Get metadata about database field (column)
Get metadata about database fields (columns)
Fetch records from the database. Always use paging when nextPageToken is provided.
List databases available in the workspace. Prefered way to search databases.
Output schema type errors in getDatabaseTool: properties use string literals ('string') instead of {type:'string'} objects. This breaks schema validation and LLM parsing.
listSpacesTool has empty inputSchema (type:object with no properties). LLM cannot infer what parameters are accepted. Should document that no parameters are required.
create_database tool has complex nested field config schema (allowedEnumValues, mirrorFields, formula) but lacks documentation of field type dependencies. E.g., which config properties apply to which field types? LLM cannot reason about valid combinations.
No error handling guidance in any tool description. LLM does not know what to do if a database is not found, if a field type is invalid, or if creation fails. Missing recovery paths.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 64 | 2025-06-18+ | v2 |
List spaces available in the workspace
Descriptions are generic and lack context for LLM selection. 'List spaces available in the workspace' does not explain when to call this vs list_databases, or what a 'space' is. Descriptions should be 50-200 chars with clear intent.