A practice MCP server implementation with Math and Weather tools using FastMCP framework
This server has critically weak tool definitions. All four tools lack parameter descriptions (every parameter has description=null), which violates the fundamental requirement that 'Every parameter needs a description explaining what it controls' (pattern:tool-description). Tool descriptions are extremely terse (10-26 characters), falling well below the 34-character p10 baseline for production tools. No input schema details beyond type declarations are visible. No output schemas are documented. No error handling guidance. The server reads like a minimal exercise, not production-grade tooling.
Add two numbers
Get the weather for a location
Multiply two numbers
Subtract two numbers
All parameters lack descriptions (description=null for every param). The rubric mandates: 'Every parameter needs a description explaining what it controls.' Parameter descriptions are required for LLM reasoning.
Tool descriptions are extremely terse (10-26 characters): 'Add two numbers' (15 chars), 'Multiply two numbers' (20 chars), 'Subtract two numbers' (20 chars), 'Get the weather for a location' (30 chars). Rubric baseline: average 194 chars, p10=34 chars. Current descriptions fall below even the p10 minimum. They lack context on WHEN to use each tool and WHAT gets returned.
No output schemas documented. The rubric requires: 'Document the output schema. LLMs need to know what fields to expect.' add/multiply/subtract return int; get_weather returns str. These should be explicitly declared in tool metadata.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 33 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 38 | - | v1 |
Parameter types lack context. For numeric tools, no mention of valid ranges, constraints, or edge cases (e.g., can 'a' and 'b' be negative? zero? What's the upper bound?). For get_weather, 'location' has no format hints (coordinates? city name?).
No error handling or recovery guidance. If get_weather is called with an invalid location or a network timeout occurs, the tool provides no actionable error message. Rubric: 'Error responses must tell the LLM what to do next.'
Math tools (add, multiply, subtract) lack semantic distinction. Three near-identical operations. Names are verb-noun (good), but descriptions are so minimal they add no value for LLM tool selection.