MCP server for OpenAI API integration
Single tool 'ask-openai' has a basic schema and minimal description. The tool name lacks a clear action verb (should be 'query-openai' or 'send-query-to-openai'). Description is only 44 characters, below the 50-200 char baseline for LLM-optimized descriptions. Parameters lack descriptions: 'query' has a 12-char description ('Ask assistant'), and 'model' has none. No output schema documented. No error handling guidance. Tool does not declare what it does beyond the bare name.
Ask my assistant models a direct question
Tool name 'ask-openai' lacks action verb. Should be 'query-openai' or 'send-query-to-openai' to clearly signal intent to LLMs.
Tool description is only 44 characters ('Ask my assistant models a direct question'). Baseline is 50-200 chars. Lacks WHAT, WHEN, and prerequisites. LLMs cannot determine when to select this tool vs alternatives.
Parameter 'model' has no description. Parameter 'query' has only 12-char description. Baseline is 72 chars per param. LLMs cannot infer what 'query' expects (free text? structured prompt? code snippet?).
No output schema documented. Tool returns TextContent but LLMs do not know what fields to expect or how to parse the response for downstream tool calls.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 38 | <=2025-11-25 | v2 |
Error handling returns raw exception text ('Error: {str(e)}') with no recovery guidance. LLMs cannot determine if error is retryable, user-fixable, or fatal.