Loading ToolBench
MCP server that integrates Twitter/X API with AI agents for creating posts
Critical gaps across all dimensions. Both tools lack parameter descriptions entirely (null values in schema). Tool descriptions are minimal (10-40 chars, below 20-char floor). No output schemas documented. No error handling guidance.
Add two numbers
Create a post on X formally known as Twitter
All parameter descriptions are null. Rubric requires non-empty descriptions for every parameter so LLMs understand what each controls. 'a' and 'b' in addTwoNumbers, 'status' in createPost are all undescribed.
Tool descriptions are too short (10-40 chars, below 20-char minimum). 'Add two numbers' and 'Create a post on X formally known as Twitter' lack context on WHEN to use, WHAT it returns, and any prerequisites. Rubric baseline: 194 chars average, 34 chars p10.
No output schemas documented. LLMs cannot plan downstream calls or extract data without knowing response structure. Both tools return {content: [{type, text}]} but this is not declared in tool metadata.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 35 | 2026-07-28+ | v2 |
createPost lacks error handling and recovery guidance. Twitter API calls can fail (auth, rate limit, invalid status). No error classification (retryable vs fatal) or actionable recovery steps for LLM.
createPost exposes destructive operation (WRITE risk) without confirmation or dry-run pattern. Agents can post to Twitter without user approval, risking unintended public statements.