Loading ToolBench
MCP server providing OpenAI GPT capabilities to Claude Code and other MCP clients using the Responses API. Features include text generation, multi-turn conversations, and server status.
Single tool 'gpt_generate' with comprehensive input schema and detailed description. Tool demonstrates good naming convention (verb_noun: 'gpt_generate'), clear parameter documentation with types and enums, and structured output schema. However, output schema documentation is embedded in description rather than formally declared, and error handling lacks structured recovery guidance. Schema completeness and parameter validation are strong, but composition patterns are limited to a single tool.
Generate text using OpenAI GPT API with a simple input prompt. This tool sends a prompt to OpenAI GPT using the Responses API and returns the generated text response. It is ideal for single-turn interactions, creative writing, code generation, analysis, and general AI assistance tasks. Args: - input (string, required): The prompt or question for GPT - model (string, optional): Model to use (defaults to GPT_MODEL env or gpt-5.4) - instructions (string, optional): System instructions for the model - reasoning_effort (string, optional): Reasoning level - none/low/medium/high - none: No reasoning (like GPT-4.1, fastest) - low: Light reasoning (fast, server default) - medium/high: Increasing reasoning depth - max_output_tokens (number, optional): Maximum output length - temperature (number, optional): Randomness 0-2 (higher = more creative) - top_p (number, optional): Top-p sampling parameter - response_format ('markdown' | 'json'): Output format (default: 'markdown') Returns: For JSON format: Structured data with schema: { "text": string, // Generated text content "model": string, // Model used for generation "usage": { "input_tokens": number, "output_tokens": number, "total_tokens": number }, "truncated": boolean // Whether response was truncated } Examples: - "Explain quantum computing in simple terms" - "Write a Python function to sort a list" - "Summarize the key points of machine learning" Note: Each call may produce different results due to model randomness.
Output schema not formally documented in tool definition. While description mentions return structure (text, model, usage, truncated), no explicit outputSchema or structured return type definition is visible in tool registration.
Error responses lack structured recovery guidance. Function handleOpenAIError() returns friendly messages but does not categorize errors as retryable/user-fixable/fatal or provide actionable next steps for LLM decision-making.
No pagination support. Tool does not accept limit/offset parameters or return pagination metadata (total_count, next_cursor), limiting handling of large generation scenarios.
Single-tool server limits composition patterns. Tool cannot chain with related tools (e.g., search_documents, summarize_text) that agents commonly invoke together.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | B | 70 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 29 | - | v1 |