MCP server for Verodat data management platform providing dataset creation, retrieval, workspace management, and AI query execution capabilities
Server has 6 tools with complete JSON Schema definitions and Zod validation. Tool names follow verb_noun convention (create-dataset, get-datasets, execute-ai-query). Descriptions are present and moderately detailed (avg ~120 chars), explaining WHAT and WHEN. However, parameter descriptions are generic ('Account ID', 'Workspace ID') without context on format, constraints, or dependencies. Output schemas are not documented, LLMs cannot predict response structure. Error handling is minimal; no recovery guidance. Security: no secrets exposed as params (good), but no permission gates or audit trail patterns visible. Composition is reasonable (6 focused tools), but filter parameters are underspecified free-form strings inviting hallucination.
Creates a new dataset with defined schema and validation rules in a specified workspace. When setting up a new data structure from scratch, storing a new type of data with specific validation rules, creating companion datasets, separating data into different logical collections, or enforcing specific data types and mandatory fields.
Executes an AI-powered query against a workspace
Retrieves AI context information for a specified workspace
Retrieves output data from a specific dataset with optional filtering and pagination
Retrieves a list of datasets from a specified workspace with optional filtering and pagination
Retrieves all workspaces for a specified account
Output schemas not documented. LLMs cannot predict response structure, forcing them to guess what fields are returned and how to chain tools. create-dataset, get-datasets, get-dataset-output, get-ai-context, execute-ai-query all lack documented return types.
Filter parameters (get-datasets, get-dataset-output) are free-form strings with no constraints or examples. LLMs will hallucinate invalid filter syntax. Should document format (e.g. 'key:value pairs', 'SQL WHERE clause', 'JSON query') or replace with structured params.
Parameter descriptions are minimal and generic. 'Account ID', 'Workspace ID', 'Dataset ID' lack context on format, range, or how to obtain them. Should include: 'Numeric account identifier (1-999999)', 'Obtain from get-workspaces response', etc.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 53 | 2026-07-28+ | v2 |
No error handling guidance. Tools lack descriptions of failure modes, recovery steps, or actionable error messages. E.g., if create-dataset fails due to duplicate name, what should the LLM do? Retry? Call get-datasets first?
No pagination guidance in tool descriptions. get-datasets and get-dataset-output accept offset/max but descriptions don't explain: What is the default limit? What is max allowed? How to detect end of results? Should return total_count or next_cursor?