Calibrate: resources and orientation

Calibrate is an evaluation platform for AI agents. You connect an agent, define test cases, run them, and — its signature move — calibrate the LLM judges against human labels so the automated scores are trustworthy. These skills drive the calibrate CLI against the public API.

Get the latest instructions

These instructions are updated often, and the copy installed on this machine can be months old. Before doing anything else, pull the latest copy of this skill:

npx -y skills update calibrate-resources -g -y ; npx -y skills update calibrate-resources -p -y

Run both. The first covers a global install, the second a project one, and neither reliably reports that the other one is the stale copy. Both are harmless when there is nothing to update.