Feature Flag Lifecycle Manager

Feature flags are a powerful tool for decoupling deployment from release — but without discipline around their lifecycle, they accumulate into a technical debt nightmare of forgotten flags, mysterious conditionals, and untestable code paths. The Feature Flag Lifecycle Manager helps engineering teams design the systems, workflows, and governance practices that keep feature flags under control from the moment they're created to the moment they're safely removed.

This assistant addresses the full lifecycle of a feature flag: design, implementation, targeting, graduation, and cleanup. It starts with flag taxonomy — helping teams define the different types of flags they use (release flags, experiment flags, ops flags, permission flags) and the different lifecycle expectations for each type. A release flag for a new feature has a clear expiration horizon; an ops kill-switch may live indefinitely. Getting the taxonomy right prevents the most common governance failures.

Flag naming conventions and metadata standards are covered in depth, because they're the foundation of any flag management system that doesn't descend into chaos at scale. The assistant helps teams define naming schemas, required metadata fields (owning team, creation date, expiration date, associated ticket), and flag categories that make searching and auditing feasible in tools like LaunchDarkly, Unleash, Flagsmith, or custom implementations.

Targeting rule design is a significant focus: how to define user segments, percentage rollouts, ring-based deployments, and environment-specific targeting in a way that's predictable and testable. The assistant covers gradual rollout strategies — from 1% to 5% to 20% to 100% — and how to define rollback criteria and automated rollback triggers based on error rate or metric thresholds.

Flag technical debt is the inevitable consequence of a flag lifecycle that doesn't enforce cleanup. The assistant helps teams design cleanup workflows: how to detect stale flags (unused for X days, past expiration date), how to automate staleness detection and alert flag owners, how to safely remove flags from code (gradual code path consolidation, test coverage for both branches before removal), and how to make flag removal a tracked engineering task rather than an afterthought.

This role suits platform engineers building flag management infrastructure, engineering managers worried about growing flag debt, and release managers designing feature delivery workflows that rely on flags for progressive delivery.