App Store Skills for Claude Code: The ASO-to-Submission Buyer's Guide

The App Store skills for Claude Code on AgentSource are five SKILL.md files that carry an iOS app from build-complete to live and earning — issued in the order you actually run them: keywords for impressions, screenshots for conversion, a paywall that survives App Review, a pre-submit gate before upload, and a review-response engine for after launch. Each is $9 one-time, or take the whole rack as the App Store Kit for $29. Every skill page shows a captured field report — real output from a documented run — before you pay, and every skill reads in Claude Code, claude.ai & Codex.

This is a ship arc, not a menu. The gear is sequenced because the jobs are: search visibility feeds conversion, conversion feeds revenue, revenue has to clear review, and review never stops after launch. Requisition in order. Here's each station on the line, what it does, and what its field report proves before you pay.

Stage 1 — App Store keywords: earn the impression

Nothing downstream matters if search never surfaces you. First requisition is the App Store Keyword Engine. The job: pull the real ranked apps for your terms, mine their tokens, and rebuild your App Name (30), Subtitle (30), and Keywords (100) from live store data instead of guesses. The field report runs it on a Productivity voice-notes app that didn't surface in the top 20 for its own brand name. It named two root causes — thin launch velocity and a wrong-category token (Journal) filing a Productivity app into Health search — then ran an intent-poisoning gate that marks each candidate term DROP or KEEP by the category its top-three results resolve to. The field report proves the skill diagnoses a real invisibility and shows its DROP/KEEP reasoning on the page, on free Apple endpoints, no paid ASO tool required.

Stage 2 — App Store screenshots: turn the impression into an install

Keywords buy the impression; the screenshot earns the tap. The App Store Screenshot Playbook is the conversion half of ASO — a 7-slot story arc, a competitor atlas, and captions short enough to get read (4–8 words, outcome before feature, verb first). The field report is an atlas pass across 7 ranked habit-tracker competitors: 6/7 used the 9:41 status bar, 0/7 showed pricing on a feature shot, and the top-converting indie broke the consensus with no device frame at all. The field report proves the skill reads a live category and tells you which conventions to copy and which single one to break — on evidence pulled from the store, not on taste.

Stage 3 — iOS paywall: bank revenue without an App Review rejection

An install that never pays is a cost, and a paywall built wrong gets the whole build bounced. The iOS Paywall Playbook is the paywall standard the factory ships on every app — the four placements, the multi-step trial, cancel-recovery — written as plain StoreKit 2 with no Superwall or RevenueCat dependency. Two documented rejections paid for it. The field report is an audit of a contract-review app: a missing session-start placement (zero grep hits) and a cancel-recovery sheet that drew a 5.6 rejection despite full disclosure — because disclosure doesn't cure the trigger; the trigger is the violation. The field report proves the skill finds both failure modes at once: the placement gap that converts at zero, and the surface that gets an app rejected.

Stage 4 — Pre-submit gate: the last check before archive and upload

Before you archive and upload, one gate stands between you and a rejection round-trip: the App Review Pre-Submit Checklist. Every check in it exists because a real submission failed. The field report shows the output format — a free-trial toggle on the paywall flagged FAIL — BLOCKER under Guideline 3.1.2(c), with the exact fix written out: drop the toggle and let the trial ride the StoreKit intro offer as subordinate text. The field report proves the checklist catches the class of rejection Apple would otherwise catch for you, and hands you the remedy, not just the flag. Run it last, right before the build goes up.

Stage 5 — Review response: work the reviews after launch

Live isn't the end of the arc. The App Review Response Engine pulls App Store and Google Play reviews, clusters them by theme and sentiment, triages the paying-and-about-to-churn ratings to the top of the queue, and drafts policy-safe replies plus a quote-backed roadmap — not a prose summary. The field report is a documented run on 14 App Store reviews: it escalated a "charged for Pro but the reports are still locked" 1★ for human sign-off, drafted a Spanish reply to a Spanish-language review, and refused to compute a 1★-spike trend on a cold start with no baseline yet. The field report proves the drafts stay honest about what's confirmed — no invented workarounds, no solicited ratings — and that the tool won't dress a metric it can't yet measure.

Buy the set — the App Store Kit

Five skills at $9 each runs $45. The App Store Kit issues the set for $29 one-time, and throws in the Deep QA Audit Loop as the build-quality station that belongs between screenshots and submission. Same field-report standard on every skill in it, same one-time price, same Claude Code / claude.ai / Codex compatibility. If you're shipping an app end-to-end rather than patching one leak, the Kit is the requisition. If you already rank fine and only need the paywall or the pre-submit gate, take the single off the shelf — the arc doesn't force you to buy stations you don't need.

Requisition by the job

The pitch here is narrow on purpose. These are the App Store skills for Claude Code that carry an iOS app the whole distance — found, installed, paying, approved, and answered — and every one of them puts its captured field report on the page before you spend a dollar. That's the standard the depot trades on: not more gear, verifiable gear. Read the report, check it does the job you came in for, then requisition. For the full ship arc in one order, take the App Store Kit; if you're still deciding which skill to run first, the Best Claude Code Skills in 2026 guide sorts the whole catalog by the job you're trying to close.