What problem does it solve?
Feature Flag Rollout prevents risky releases from turning into production incidents by forcing a complete rollout plan (kill-switch/percentage/cohort), explicit success metrics, and testable rollback criteria before any durable work proceeds.
Core Features & Use Cases
- Rollout shape selection: Chooses kill-switch, percentage, cohort, or dark-launch based on worst plausible outcomes and audience targeting needs.
- Rollback contract + gates: Defines numeric SLI thresholds, a wall-clock rollback budget, and qualitative tripwires for incident-grade failures.
- Staged execution planning: Specifies stage steps with soak times and advance gates, plus a cleanup deadline to prevent long-lived flag debt.
- Evidence-first handoff: Blocks completion when required source evidence, owner approval, validator output, or runtime receipt is missing, returning a structured BLOCKED response.
Quick Start
Ask your AI to plan a staged rollout for a new risky behavior with a kill-switch and measurable rollback gates using feature-flag-rollout.