Loading ToolBench
Try your luck with AI: randomized tarot, dice, and coin results for creative storytelling and TRPG adventures via CLI and MCP.
Three tools with complete schemas and reasonable descriptions. Naming is clear and action-oriented (dice, coin, tarot). Descriptions are adequate (100-200 chars) but lack guidance on when to use each tool or dependencies. All parameters have types and descriptions. No error handling guidance or output schema documentation. Tools are read-only and well-scoped, but descriptions could be more LLM-optimized with explicit use-case hints.
Flip a coin one or more times. Returns a JSON array of heads/tails strings, or true/false strings when boolean=true.
Roll dice one or more times. Returns a JSON array of roll results normally, or an object {"rolls": [...], "total": N} when sum=true or modifier is non-zero. Supports dice notation (e.g. "3d10", "d6+2") or explicit count/sides/modifier parameters.
Draw one or more Major Arcana tarot cards. Returns a JSON array of objects with `card` (card name) and `orientation` ("upright" or "reversed") fields. Card names use snake_case by default (e.g. "the_fool"), or proper case when case="proper".
No output schema documentation. Tool descriptions state what is returned (JSON array, objects with fields) but lack formal schema definitions. LLMs cannot reliably parse or chain results without explicit output structure.
Descriptions lack use-case guidance. No explanation of WHEN to call each tool or how they differ from each other. LLMs may struggle to select the right tool when multiple randomization options exist.
No error handling guidance. Descriptions do not explain failure modes (e.g., invalid notation in dice, out-of-range count in tarot). LLMs have no recovery path if a call fails.
Parameter descriptions contain example values (e.g., '3d10', 'd6+2', 'the_fool'). LLMs may reuse these literally instead of adapting to context.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 60 | 2026-07-28+ | v2 |
Tarot tool has implicit max constraint (max: 22) documented in description but not enforced in schema. Schema should include maxItems or explicit constraint.