Loading ToolBench
MCP server that converts React code to video using Remotion library
Single tool with minimal description (4 words, well below 20-char threshold). Input schema present with proper types and required fields, but parameter descriptions are terse (avg 20 chars). No output schema documented. No error handling guidance. Tool name 'react_code_to_video' lacks action verb prefix (should be 'convert_react_code_to_video' or 'render_react_to_video'). No indication of side effects (WRITE risk) in description. Composition is monolithic, bundling, rendering, and file I/O are internal but not exposed for agent control.
convert react code to video
Tool description is 4 words ('convert react code to video'), far below 20-char minimum. Provides no context on when to use, prerequisites, or what is returned. LLM cannot determine if this is the right tool or what to expect.
No output schema documented. Tool returns a file path string, but LLM has no way to know the format, whether it's a local path or URL, or how to use it downstream. Breaks tool chaining.
Parameter descriptions are minimal (avg 20 chars). 'react code using remotion library' assumes LLM knows Remotion syntax. 'width of the video' lacks constraints (min/max pixels). 'duration of the video' does not specify units (seconds? milliseconds?). Code inspection shows duration is treated as milliseconds, but description does not state this.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 36 | 2026-07-28+ | v2 |
Tool name lacks action verb. 'react_code_to_video' is a noun phrase; should start with verb like 'convert_', 'render_', or 'generate_'. LLM cannot infer intent from name alone.
No error handling guidance. Tool can fail if Remotion bundling/rendering fails, but description provides no recovery hints. LLM will not know whether to retry, ask user for different code, or escalate.