Loading ToolBench
Two tools with minimal documentation. 'parentify' has a basic description (47 chars) and proper schema with typed parameter. 'touch' has a vague description (54 chars) lacking context on what 'hello doc' is or when to use it. Neither tool has descriptions explaining WHAT they do, WHEN to use them, or WHAT they return. Parameter descriptions are missing entirely. No output schemas documented. No error handling guidance. Naming is non-standard: 'parentify' lacks verb_noun clarity; 'touch' is generic and does not convey the side effect (marking as updated). No tool annotations (readOnlyHint/destructiveHint) despite 'touch' being a WRITE operation.
Parentify a text: wraps a text within parenthesis.
Marks the hello doc as updated, notifying any subscriber
Missing output schemas for both tools. LLMs cannot plan downstream calls or extract results without knowing what fields are returned.
No parameter descriptions. 'text' parameter in parentify lacks guidance on format, length, or constraints. LLMs cannot infer parameter meaning from names alone.
Generic tool names lack action verbs. 'parentify' is non-standard (should be 'wrap_text_in_parentheses' or similar). 'touch' is vague and does not convey the side effect.
Missing tool annotations. 'touch' is marked WRITE but has no destructiveHint annotation. 'parentify' is READ_ONLY but has no readOnlyHint. Agents cannot determine safety/idempotency.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 32 | 2026-07-28+ | v2 |
Vague descriptions lack context. 'touch' description ('Marks the hello doc as updated, notifying any subscriber') does not explain what 'hello doc' is, when to call it, or what it returns. No recovery guidance on errors.