Loading ToolBench
Dead-simple MCP server for vision analysis with Google Gemini Flash-Lite
Two vision analysis tools with clear, actionable descriptions (120-180 chars each). Input schemas present with proper types and descriptions for all parameters. However, output schemas are completely undocumented, LLMs cannot plan downstream operations or extract structured data. No error handling guidance, no parameter constraints (e.g., max 10 images), and no security/permission documentation. Tool names are verb-noun compliant (analyze_image, analyze_images), but the dual-tool design lacks clear differentiation guidance for LLM selection.
Analyze an image using Google Gemini Flash-Lite vision model. Supports URLs, file paths, file:// URLs, and base64 data URIs. Returns text analysis of the image content.
Analyze multiple images together using Google Gemini Flash-Lite vision model. Supports up to 10 images. Accepts URLs, file paths, file:// URLs, and base64 data URIs. Returns text analysis comparing or analyzing all provided images.
Output schema completely undocumented. LLMs cannot infer what fields the response contains, preventing structured downstream processing and forcing unstructured text parsing.
No error handling guidance. Errors like 'Invalid data URI format', 'Image too large', 'Failed to fetch image' are returned raw without recovery hints. LLMs cannot self-correct or know whether to retry.
Parameter constraints not documented in descriptions. 'Maximum 10 images' is enforced in code but not stated in the analyze_images parameter description. LLMs cannot know the limit and may attempt invalid calls.
No guidance on when to use analyze_image vs analyze_images. LLMs must infer from names alone. Add: 'Use for single image analysis' and 'Use to compare or analyze multiple images together' to descriptions.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | 2026-07-28+ | v2 |
API key (GEMINI_API_KEY) is correctly injected server-side, but no documentation of required permissions or scopes. Add: 'Requires GEMINI_API_KEY with Gemini API access' to tool descriptions.