Chat interface designed for developers to test and interact with MCP
This server exhibits pervasive quality gaps across all dimensions. Five tools are defined, but most lack adequate descriptions and parameter documentation. Input schemas are present for only 3 of 5 tools; the others (getBooks, get_date, get_products, get_company_info) have empty required arrays with no documented parameters, even though some appear to accept arguments. Descriptions are present but generic and minimal (10-50 chars for most tools), falling below the 50-200 char baseline for LLM-optimized descriptions. Parameter descriptions are almost entirely absent, e.g., get_time_for_user has a 'name' parameter with description 'The name of the account holder' (35 chars), which is barely adequate but lacks guidance on format, constraints, or what happens when the parameter is absent. No output schemas are documented. Error handling is minimal, no guidance on retryability, recovery steps, or what happens when a tool fails. The tool definitions appear to be registered via fastmcp decorators (@mcp.tool()), but explicit schema validation rules, enums, and constraints are absent. The JS books_server.js registers its tool via setRequestHandler, but still lacks output schema documentation. Overall, this is a bare-minimum implementation, definitions exist but are skeletal and would confuse LLMs about when and how to invoke each tool.
Returns a list of books
Returns information about VSL PVC company.
Returns the current date and time.
Returns the list of products offered by VSL PVC company.
Get time for user.
Missing or trivial tool descriptions. 4 of 5 tools have descriptions under 50 characters (getBooks: 27 chars, get_date: 35 chars, get_products: 58 chars, get_company_info: 50 chars). LLMs cannot reliably disambiguate tools or infer selection criteria from descriptions this brief. Rubric baseline: 50-200 chars for adequate descriptions.
Input schemas lack parameter descriptions and constraints. get_date and get_products declare empty parameter objects with no required array and no documentation of what they return. get_time_for_user has a 'name' parameter but provides no guidance on constraints (e.g., must match 'Alex' based on code inspection), valid formats, or error conditions. get_books and get_company_info have identical issues. Rubric baseline: 100% of A+ tools document ALL parameters.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 37 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 37 | - | v1 |
Output schemas are not documented. None of the 5 tools include documentation of what fields are returned, types, or structure. Code inspection shows tools return JSON strings (books array, datetime, products list, company info object), but the tool definitions do not declare these schemas. LLMs cannot plan downstream operations or extract the right data without documented output schemas.
No error handling guidance. Tools do not declare what happens on failure (e.g., when get_time_for_user receives a name other than 'Alex', it returns a string 'Need to provide user name...' rather than a structured error). No error recovery paths, retryability classification, or actionable error messages are provided. Rubric baseline: 100% of A+ tools guide error recovery.
Missing input validation constraints. get_time_for_user silently returns an error message if name != 'Alex', but this constraint is not declared in the schema. No enums, regex patterns, min/max bounds, or format constraints are specified. LLMs will invoke tools with invalid inputs (e.g., any name) without guidance on what values are acceptable.