What problem does it solve? Deploying new features directly to all users couples deployment with release, making risky launches, A/B tests, and instant rollbacks difficult. This Skill decouples deployment from release by setting up a complete feature flag system with naming conventions, rollout stages, and enforced cleanup. ## Core Features & Use Cases - Flag Registry and Naming Convention: Produces a registry at wiki/guides/feature-flag-registry.md with typed prefixes (release-, exp-, ops-, permission-), owners, and expiry dates. - OpenFeature SDK Setup: Generates vendor-neutral flag evaluation code for TypeScript/Node.js, Python, and Go that works with LaunchDarkly, GrowthBook, or environment variable backends. - Progressive Rollout Pattern: Implements consistent-hash percentage rollouts (1% → 5% → 25% → 50% → 100%) plus a CI hygiene check that blocks merges when flags expire. - Use Case: When shipping a new checkout flow, use this Skill to wrap it in a release-checkout-v2 flag, roll it out to 1% of users, monitor error rates at each stage, and enforce code cleanup within two sprints of reaching 100%. ## Quick Start Ask the AI to set up a feature flag with gradual rollout for the new feature you are about to deploy.