Loading ToolBench
A simple API for fetching Rick and Morty characters
Two tools with basic schemas and minimal descriptions. Both tools have input schemas with type constraints (number, min/max), but descriptions are terse (10-40 chars) and lack context for LLM selection. No output schema documented. No error recovery guidance. Parameters lack detailed format/constraint descriptions beyond JSON Schema. No parameter descriptions for get-characters' pageSize (mentions 'default: 20, max: 20' in description but not in param annotation). Naming is clear (verb_noun pattern), but overall quality is below production baseline (194 chars avg description, 72 chars avg param annotation).
Fetch a character by ID
Fetch characters with pagination
Tool descriptions are too brief (10-40 chars vs 194 char baseline). 'Fetch a character by ID' and 'Fetch characters with pagination' lack context for when to use each tool, what they return, or dependencies.
No output schema documented. LLMs cannot plan downstream calls or extract fields. Responses are formatted as plain text strings, not structured objects with typed fields.
Error handling returns generic text ('Error fetching character with ID X') without recovery guidance. No indication whether error is retryable, user-fixable, or fatal.
Parameter descriptions incomplete. 'page' and 'pageSize' in get-characters have descriptions in Zod .describe() but lack detail on valid ranges, defaults, and constraints. Baseline is 72 chars per param; these are ~50 chars.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | 2026-07-28+ | v2 |
No pagination metadata returned. get-characters returns formatted text but omits total count, next_cursor, or page info needed for multi-page workflows.