Loading ToolBench
MCP server that exposes content generation skills (playbooks) for video, image, and 3D model generation. Provides skill listing and retrieval via embedded catalog.
Two tools with minimal definitions. list_skills has a basic description (56 chars) but no input schema. get_skill has a description (67 chars) and a single parameter (skill_id: string) with minimal description (47 chars). Neither tool documents output schema, error handling, or recovery guidance. No parameter constraints (enums, ranges), no pagination support despite likely returning lists. Descriptions lack WHEN-to-use context and prerequisites. No tool annotations (readOnlyHint, idempotentHint). Tool names follow verb_noun pattern (acceptable) but descriptions are below production baseline (194 chars avg). Schema completeness is poor, only get_skill shows any input structure; list_skills has none visible.
Retrieve detailed information about a specific skill by ID
List all available content generation skills from the catalog
list_skills has no input schema visible. Cannot verify pagination, filtering, or sorting parameters. Likely returns unbounded list without limit/offset.
No output schemas documented for either tool. LLMs cannot plan downstream calls or extract required fields (e.g., skill_id for chaining). Violates pattern:tool requirement.
get_skill parameter 'skill_id' description is 47 chars, below minimum. Does not explain format, constraints, or how to obtain a valid ID. LLMs cannot infer whether to pass UUID, slug, or integer.
No error handling guidance. What happens if skill_id is invalid? Does list_skills return empty or error? No recovery hints for LLM.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 40 | 2026-07-28+ | v2 |
Descriptions lack WHEN-to-use context. 'List all available content generation skills' does not explain when to call this vs get_skill, or what structure is returned.