Loading ToolBench
MCP server providing utility tools for arithmetic and data fetching
Both tools have minimal descriptions (under 50 chars) and parameter descriptions are entirely missing (null). The 'add' tool is trivial; 'fetch-data' makes external API calls but lacks error context, retry guidance, or output schema documentation. No tool annotations (readOnlyHint, destructiveHint). Schemas are present via Zod but parameter descriptions are absent, violating the 100% baseline for A+ tools. No composition guidance or chaining IDs documented.
Add two numbers
fetch data from api by pageSize
All parameter descriptions are null. LLMs cannot infer what 'a', 'b', or 'pageSize' control without explicit text. Violates pattern:tool-description baseline (100% of A+ tools have param descriptions).
Tool descriptions are under 50 characters and lack WHAT/WHEN/WHY context. 'Add two numbers' and 'fetch data from api by pageSize' do not explain when to use them or what they return. Baseline: 194 chars average for A+ tools.
fetch-data makes external API calls but provides no error recovery guidance. Errors are caught and returned as text, but LLM has no hint to retry, check network, or call an alternative. Violates pattern:recovery-guide.
No output schema documented. LLMs cannot plan downstream tool calls or extract structured data. 'add' returns a text string; 'fetch-data' returns an array of text items. Neither is formally specified.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 35 | 2026-07-28+ | v2 |
fetch-data lacks pagination support. If the API returns 100+ articles, all are returned in one response, risking context window exhaustion. No limit parameter or next_cursor documented.