Configure and interpret LaunchDarkly AI Config online evaluations with judge attachments

Attach judges to LaunchDarkly AI Config variations, create custom judges, set sampling rates, and interpret production quality signals from online evaluations.

Prerequisites

LaunchDarkly AI Configs; LaunchDarkly API token or SDK access; custom skill-capable agent client

Installation

Use the upstream install or setup path that matches your environment:

• git clone https://github.com/launchdarkly/ai-tooling.git
• npx skills add launchdarkly/ai-tooling
• npx skills add <owner/repo>

Basic usage or getting-started notes:

• This repo is a Claude Code plugin. Installing it gives you all the skills above plus the LaunchDarkly MCP server.

• Open Claude Code and run /plugin install.

• Search for LaunchDarkly, or install directly from the repo URL:

• Source: https://github.com/launchdarkly/ai-tooling/tree/main/skills/ai-configs/aiconfig-online-evals

• Extracted from upstream docs: https://raw.githubusercontent.com/launchdarkly/ai-tooling/HEAD/README.md

Documentation

• https://docs.launchdarkly.com/home/ai-configs/online-evaluations

Source

• Agent Skill Exchange