Launchdarkly Experiment Setup

Walks you through the full lifecycle of A/B tests in LaunchDarkly: creating experiments with metrics and treatments, starting iterations to collect data, and stopping with a declared winner. The skill handles the finicky bits like ensuring metrics exist first, grabbing the right ruleId and flagConfigVersion from your feature flag, and knowing which fields you can edit while an experiment is running versus when you need to stop and start a new iteration. If you're running experiments through LaunchDarkly rather than just toggling flags, this gives Claude the context to set them up correctly without you needing to remember the API's state machine rules.

Install to Claude Code

npx -y skills add launchdarkly/agent-skills --skill launchdarkly-experiment-setup --agent claude-code