An MCP client that uses a DSP agent to interact with MCP servers and execute tools
This MCP server has 6 tools with basic definitions but significant quality gaps. Tool names follow verb_noun convention adequately (reverse_text, calculator_sum, count_words), but descriptions are inconsistent and shallow. Five tools have functional JSON schemas with proper type definitions, but parameter descriptions are minimal. Tool #5 (count_characters) has a copy-paste error in its description ('Count the number of words' instead of characters). Error handling is absent, tools return raw Python values with no guidance for recovery. No input validation, no pagination support, no structured output documentation, and no security considerations. The server is a demonstration/testing tool rather than production-grade, evidenced by the commented-out execute_python_code tool and the 'fake-server' naming.
Calculate the multiplication of a list of numbers (e.g., 2 * 3 * 4 = 24).
Calculate the sum of a list of numbers.
Count the number of words in a text.
Count the number of words in a text.
List available tools from all connected MCP servers.
Reverse the input text.
Description copy-paste error in count_characters tool. Docstring says 'Count the number of words in a text' but the tool counts characters. This misleads LLMs about what the tool does.
Insufficient parameter descriptions. All six tools have parameters with minimal or generic descriptions ('The text to reverse', 'List of numbers to sum'). Descriptions should clarify expected format, range, constraints, and when to use the tool. Average param description length is ~15 - 20 chars; baseline for A+ tools is 72+ chars.
No output schema documentation. Tools return simple types (string, float, int, array) without documenting the structure or fields. LLMs cannot plan downstream tool chains or validate returned data.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 50 | 2026-07-28+ | v2 |
| 2026-03-09 | D | 50 | - | v1 |
No error handling or recovery guidance. Tools return raw Python exceptions without actionable error messages. An LLM receiving 'TypeError: unsupported operand type' has no idea how to correct the input or proceed.
list_tools tool has no documented schema. Input is an empty object (no parameters), but the output schema is not visible. What fields does it return? How is the list structured?
Tool descriptions are too brief (10 - 35 characters). Baselines for A+ tools are 50 - 200 chars. Descriptions should state WHAT the tool does, WHEN to use it (vs similar tools), and WHAT it returns. Current descriptions are too generic for LLM selection.