Flag Rollout Planner
Plan one feature's flagged launch: flag spec, staged ramp, guardrail metrics, auto-rollback thresholds, and an expiry date.
Before & after
Same prompt, same agent — the only difference is the skill.
Feature flags are useful for controlling the rollout of new features. You can gradually release a feature to a percentage of users and monitor for issues before going fully live. Some teams use feature flag services that let you toggle features on and off remotely without redeploying code. You'd want to think about who your target users are and how to measure whether the feature is working as intended. It's also important to clean up old flags after a feature is fully rolled out. Different flag services have different capabilities and pricing models.
FEATURE FLAG ROLLOUT ARCHITECTURE FOR BILLING V2: FLAG NAMING STRATEGY • billing_v2_overhaul (main kill switch) • billing_v2_ui_redesign (UX changes only) • billing_v2_invoice_calc (calculation logic) • billing_v2_reporting (new reports) ROLLOUT SCHEDULE (10percent weekly) Week 1: Canary segment (internal QA + 10 friendly accounts) Week 2: 10percent random sample (monitor error rates, latency) Week 3: 20percent (expand if no issues) Week 4: 50percent Week 5+: 100percent (deprecate flag, remove from codebase) TARGET AUDIENCE DEFINITION • Include only accounts with Billing Admin role • Exclude accounts with open billing disputes • Exclude accounts in onboarding (first 7 days) • Include accounts in trial and paid tiers equally MONITORING & ROLLBACK CRITERIA RED FLAGS (automatic rollback): • Error rate spikes above 5percent on invoice creation • P99 latency exceeds 2 seconds • Payment processing fails more than 1percent GREEN LIGHTS (proceed to next segment): • Error rate stable at <1percent • User feedback score >4.0 in new UI • Zero billing dispute increases CLEANUP PLAN After 30 days at 100percent, flag removal: 1. Delete feature flag config from system 2. Remove if/else branches from codebase 3. Archive flag history and ruleset
About this skill
Before a risky change goes to production, get a concrete rollout plan: exactly how the flag is defined, how traffic ramps, and which numbers trigger a rollback. You get one flag per launch with an owner and an expiry date from day one.
What it does
- Specifies the flag: key, type, variants, fallback value, bucketing key, evaluation point, owner, and removal date
- Builds a staged ramp (internal, 1%, 10%, 50%, 100%) with hold times and go criteria sized to your traffic
- Defines 3-5 guardrail metrics compared across flag-on and flag-off cohorts, with rollback thresholds
- Checks rollback safety, including in-flight users and data written while the flag was on
What you get
A rollout plan with the flag spec, ramp table, guardrails, rollback procedure, pre-launch test matrix, and expiry ticket.
Best for
Engineers and release managers shipping payment, pricing, data-path, or other high-blast-radius changes.
SKILL.md preview
---
name: flag-rollout-planner
description: Write the flag spec and staged rollout plan for one feature about to ship behind a feature flag, including guardrail metrics, automatic rollback thresholds, kill switch behavior, and an expiry date. Use when the user says "ship this behind a flag", "plan the rollout", "canary this feature", "percentage rollout", "what should the kill switch do", or is about to launch a risky change to production, including casual asks like "should this be behind a flag?", "what percentage should we start at", or "when do we roll back the new checkout".
---
# Flag Rollout Planner
… (sign up to view the full skill)More development skills
View all Development skills →Chaos Experiment and Game Day Planner
Plans chaos experiments and game days: hypothesis, blast radius, abort conditions, tool manifests, rollback, and post-run review.
Claude Code Cost and Latency Routing
Configure Claude Code Router with model tiers, explicit fallbacks, compatibility checks, and a cost-aware validation plan.
CI/CD Pipeline Builder with Gated Deploys
Generates CI/CD workflows from your repo's real stack: cached fast-fail checks, pinned actions, and staging-to-prod deploys with gates.
Terraform Modules and Safe Deployments
Get reusable Terraform modules, safe state migrations, explicit provider wiring, and CI gates tied to reviewed plans.