Tool Safety

What tool descriptions in the index tell the model to do.

A tool description is text that enters the model's context and is read as instruction, so it is the one part of a third-party server that acts on an agent before the agent calls anything.

The measurement is published because the answer is a surprise: the corpus is close to clean. It is deliberately NOT a score, a badge or a ranking signal. Findings are counts and excerpts for a human to read, and the classes the detectors refuse to match are published alongside the ones they do, because an earlier draft of this scan was wrong about four legitimate publishers.

get/tool-safety

Request

  • Base URL: https://neuronto.com
  • URL: https://neuronto.com/tool-safety
  • Auth: none declared

Query parameters

entrystring

Response

Successful Response

unknown required

Changes

No recorded changes to this operation across all 1 revision of this API.