A skill is a folder of instructions and scripts that Claude Code loads when the task fits, like “get this build into TestFlight” or “is this ready to ship”. We wrote these for our own release work and run them on every build.
Paid
Store PreflightPaid
One command before every App Store or Play build. Catches a release bundle signed with the debug key, vague permission strings, a missing export compliance flag, Google sign-in with no Sign in with Apple, and an ATT prompt App Review can't reach.
Store ScreenshotsPaid
Drives your real app on simulators and emulators, captions every screen in every listing language, and checks each size the stores accept. Also makes the Play preview video and the App Preview.
App CompliancePaid
Audits AdMob consent, the in-app link to change it, sample or sibling ad ids, child-directed ad settings and the iOS privacy manifest, then drafts the privacy label and Data safety answers.
Launch BrandingPaid
The launcher icon, adaptive and themed icons, the native splash and the animation into the first screen, drawn in code and checked on every shape.
Bug HuntPaid
Finds what a green test suite still ships: wrong data on screen, saves that no longer load, layouts that break at large text. Each fix comes with a test that fails first.
iOS Release to TestFlightPaid
From a fresh fork with no records to a build in review, with the failures we hit written out, down to the price dropdown that won't accept the word "Free".
New App from a ForkPaid
Cut a new app from one you already have, rebrand it fully, and carry it through the release checks.
Release CadencePaid
For an account with many apps: which one takes today's update slot, the gates it runs first, and submission only after you say yes.
Four of them (iOS Release, New App, Bug Hunt and Release Cadence) are sold together as the Flutter Indie Release Pipeline. Prices are on Agensi.
Free
app-similarityFree
Measures how much an app shares with its siblings (source, listing text, screenshots, bundled assets) before you submit. Written after a 4.3 rejection.
listing-accuracyFree
Checks that every claim in the store listing is something the app actually does.
niche-scoutFree
Measures an app idea against Apple's search API before any code: rating counts, staleness, and the terms nobody owns.
play-releaseFree
The Play Console declarations in the order Play wants them, an .aab preflight, and which steps only a human can do.
polishFree
The last 10%: accessibility, large text, tablets, safe areas, and the things that make an app feel unfinished.
perfFree
Finds work repeated every frame or every render, and measures whether the fix helped.
The free ones are MIT-licensed, in one GitHub repo. For what app-similarity finds on real apps, see how much of a fork is still its parent, where we ran it on 29 of ours.