An AWS Lambda-hosted MCP server deployed via CDK with API Gateway integration
This server has a single tool 'add' with critical definition gaps. The tool is registered with a Zod schema (a: z.number(), b: z.number()) but has no description at all, and the parameters lack descriptions. The naming is generic and non-idiomatic (should be verb_noun style like 'add_numbers' or 'compute_sum'). No output schema is documented. This is a minimal proof-of-concept implementation that does not meet production quality baselines.
Tool 'add' has no description. LLMs cannot determine when to call it or how it differs from similar arithmetic tools. Description field is entirely missing.
Parameters 'a' and 'b' have no descriptions. LLMs rely on descriptions to understand what each parameter controls. 'a' and 'b' are mathematically clear to humans but do not explain business context or constraints.
Tool name 'add' is generic and does not follow verb_noun naming convention. Should be 'add_numbers', 'compute_sum', or 'calculate_addition' to clearly signal intent and distinguish from other arithmetic operations.
Output schema is not documented. The tool returns {content: [{type: 'text', text: String}]} but LLMs have no way to know the structure or extract the numeric result. No schema definition visible for return type.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 32 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 0 | - | v1 |
No parameter constraints defined. Parameters a and b are unbounded, LLMs could pass Infinity, NaN, or extremely large numbers. Should specify minimum/maximum ranges if applicable.
No error handling or recovery guidance. If invalid input is passed (non-numeric), tool fails silently or returns a cryptic error. No guidance for LLM to self-correct.