Loading ToolBench
A custom MCP server that provides super secret combinations
Scoring was not performed
Tool description is only 18 characters and cryptic ('What\s the password?'). Baselines require 34+ characters with clear WHAT/WHEN/WHY guidance. This description does not help LLMs decide when to call the tool.
Output schema is not documented. The tool returns a string (state + soup combination), but LLMs have no way to know this from the tool definition. Pattern:tool requires output schema documentation.
Tool description contains a typo or grammatical error ('What\s the password?' with backslash-escaped apostrophe instead of proper quote). This signals poor tooling quality and confuses LLM parsing.
No error handling documentation. The tool handler contains no try-catch logic visible for error conditions. If random selection fails or the handler throws, there is no guidance for the LLM on recovery.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 17 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 28 | 2024-11-05+ | v1 |
Tool appears to be a novelty/demo function (random state + soup selection) rather than a production capability. No real-world use case is apparent, and the implementation does not align with serious agent tool patterns.