FastMCP template server with tools, resources, prompts, subscriptions, and authentication
This MCP server has 4 tools with basic definitions but significant quality gaps. Tool names are action-oriented (echo, hello_tool, raise_error, list_demo_capabilities) but descriptions are minimal and lack depth. Input schemas are present but sparse, most parameters lack detailed descriptions. Output schemas are entirely undocumented. Error handling is absent (raise_error tool deliberately throws an exception with no recovery guidance). No tool demonstrates the LLM-optimized description length baseline of 50-200 characters. No parameters have constraints (enums, ranges, patterns). The server reads as a template/demo rather than production-grade. Average tool score: 42/100.
Echoes back the provided text.
Demo tool that returns a greeting.
Lists all registered capabilities.
Tool that always raises an error for testing.
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields.
Descriptions are too short (avg 38 chars vs baseline 194 chars). Missing WHAT/WHEN/WHY context that LLMs need for tool selection.
Parameter descriptions are missing or trivial (avg 13 chars, all below 20 char threshold). 'Text to echo' and 'Name to greet' are ambiguous without format/range context.
raise_error tool provides no recovery guidance. Exception is thrown with no actionable error message for LLM retry.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 48 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 26 | - | v1 |
No input constraints (enums, min/max, patterns, regex). 'name' parameter in hello_tool has no length limit; 'text' in echo has no size constraint.
list_demo_capabilities output schema is unknown. LLMs cannot reason about the structure, field names, or downstream usage.