Loading ToolBench
A unified Model Context Protocol server for MiniMax CLI (mmx)
8 tools with decent naming (verb_resource pattern) and solid descriptions (150-250 chars average). All tools have complete input schemas with typed parameters and descriptions. However, output schemas are not documented, the server returns raw API responses without specifying structure. Parameter constraints are underutilized (e.g., aspect_ratio and resolution accept free-form strings instead of enums). Error handling guidance is minimal, tools provide no recovery hints on failure. Security is a concern: API keys are accepted as command-line arguments and passed to spawned child processes, risking exposure in process listings.
Generate images with MiniMax. If no output path is provided, images are saved to the default output directory (e.g. ./mmx_image) and their file paths are returned.
Generate music with MiniMax. Supports instrumental or vocal music. If lyrics and style are provided, generates a song. If no output path is provided, the audio file is saved to the default output directory (e.g. ./mmx_music).
Show MiniMax usage quotas and remaining limits for each model or service.
Search the web using MiniMax. Returns search results with titles, snippets, and URLs. Use this when you need up-to-date information, facts, news, or references from the internet.
Synthesize speech with MiniMax TTS. Converts text into an audio file. If no output path is provided, the file is saved to the default output directory (e.g. ./mmx_speech).
No output schemas documented. Server calls 'mmx' CLI and returns raw JSON; LLMs cannot infer result structure or plan downstream operations. For example, mmx_search likely returns a list of results, but keys, nesting, and pagination are undocumented.
Parameter constraints not enforced via schema. 'aspect_ratio' (mmx_image_generate) accepts free-form strings instead of enum ['16:9', '1:1', '4:3']. Same for 'resolution' (mmx_video_generate) and 'style' (mmx_music_generate). LLMs will hallucinate invalid values.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 61 | 2026-07-28+ | v2 |
Chat with MiniMax text models. Supports system prompts and optional JSON mode. Use this for general reasoning, writing, coding help, or structured output.
Generate video with MiniMax from a text prompt. If no output path is provided, the video is saved to the default output directory (e.g. ./mmx_video) and its file path is returned.
Analyze an image with a custom instruction using MiniMax Vision. Useful for: UI/UX critique, extracting text, identifying objects, evaluating designs, or answering specific questions about the image.
API key exposure risk. Line 11-16 shows API key is read from command-line args or environment and passed to spawn('mmx', ...). Command-line arguments are visible in process listings (ps aux). This violates secret-injection pattern, credentials should never appear in tool parameters or process arguments.
No error recovery guidance. Tools silently fail if 'mmx' CLI is not installed, if API key is invalid, if rate limits are hit, or if output parsing fails. Error messages should guide LLM on retryability and next steps, e.g., 'API key invalid. Reconfigure MCP server with MINIMAX_API_KEY environment variable.'
No pagination for list tools. mmx_search likely returns many results; no limit, offset, or page parameters. High-volume results blow context windows. Add parameters: limit (default 20), offset (default 0), and return a total_count field.
mmx_quota_show lacks description depth. Description is 62 characters, too brief. Expand to: 'Retrieve your current usage quotas and remaining limits for each MiniMax model or API service. Useful before generating large batches of content (images, videos, audio) to avoid hitting limits. Returns available quotas, used amounts, and reset schedules.'
Optional parameters lack guidance on interactions. For example, mmx_music_generate accepts both 'prompt', 'lyrics', and 'style'. Are all three required? Can you omit prompt if lyrics are provided? Document: 'prompt is required. lyrics and style are optional; if both provided, generates a vocal track in the specified style.'