Loading ToolBench
MCP server for generating images using OpenAI's DALL-E API and Stable Diffusion
Single tool 'generate-image' has a clear verb-noun name and reasonable schema with typed parameters (string, number). However, the description is generic (67 chars, below the 194-char baseline), lacks WHEN/WHY guidance, and omits critical details about prerequisites (API key setup, rate limits, cost). Parameters lack descriptions for 'destination' (empty string). No output schema documented. No error recovery guidance. Tool modifies state (writes files) but description does not explicitly state this. Missing idempotency and confirmation patterns for a destructive operation.
Generate an image using Stable Diffusion based on a text prompt
Tool description is too short (67 chars) and lacks actionable context. Does not explain prerequisites (API key, rate limits, cost), when to use it vs alternatives, or what happens on failure.
No output schema documented. LLM cannot predict the structure of the response (text content, file path, URL) and cannot plan downstream operations.
Parameter 'destination' has no description. LLM cannot determine if it expects an absolute path, relative path, filename only, or directory.
Tool performs file I/O (writes to disk) but description does not state this is a state-modifying operation. LLMs need explicit 'creates', 'modifies', or 'deletes' language to understand side effects.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 46 | <=2025-11-25 | v2 |
No error recovery guidance. Code returns InternalError for API failures, network errors, and file I/O errors, but does not tell the LLM what to do next (retry, check API key, free disk space, etc.).