App Store

App Store

One skill for App Store discoverability, submission readiness, and codebase review. It operates in three modes. Pick the mode from the request, then follow that mode's steps. Deep detail lives in references/, loaded at the point of use.

Mode Routing

| If the request is about... | Mode | Read | |---|---|---| | Ranking, discoverability, conversion, keywords, app name/subtitle, description copy, screenshots/captions, Custom Product Pages, In-App Events, A/B tests, ratings strategy, review-prompt timing, localized metadata | OPTIMIZE | references/aso-keywords.md, references/aso-product-pages.md | | Getting through App Review: rejection reasons, guidelines, privacy manifests, required-reason APIs, ATT, IAP/StoreKit rules, entitlements, metadata field limits, screenshot sizes, submission workflow, appeals | PREPARE | references/review-checklists.md, references/review-privacy-manifest.md | | "Review my app/codebase for rejection risks", "audit before submission", walking actual source + config + metadata like a reviewer | AUDIT | references/audit-procedure.md |

A single request can span modes (e.g., "improve keywords but don't get rejected" = OPTIMIZE with a PREPARE cross-check). Run the primary mode and pull the specific fact from the other mode's reference rather than switching wholesale. Metadata field limits and screenshot sizes are owned by PREPARE (references/review-checklists.md); OPTIMIZE works within them.

Read First

Before advising, locate what actually exists. Do not generate advice against an imagined app.

• OPTIMIZE: read the current listing metadata (app name, subtitle, keyword field, description, promotional text, screenshots/captions, category) and any App Analytics or Search Ads data the user has. Advise against the real copy, not a blank slate.
• PREPARE: locate the app target, Info.plist, *.entitlements, PrivacyInfo.xcprivacy, StoreKit configuration/usage, and existing App Store Connect metadata before judging readiness.
• AUDIT: follow the read-first inventory in references/audit-procedure.md (Step 2) -- app target, Info.plist, entitlements, privacy manifest, StoreKit usage, permission strings, login/paywall flows, metadata.

---

Mode: OPTIMIZE (App Store Optimization)

Improve search visibility and conversion. Separate the two pillars explicitly in any plan.

Visibility (search/browse): app name, subtitle, keyword field, primary category, localization, ratings/reviews, In-App Events, Custom Product Pages with assigned keyword search visibility. Conversion (product-page decision path): screenshots, app previews, description, promotional text, Custom Product Page messaging, Product Page Optimization tests, and alignment between public claims, screenshots, and the real in-app UI.

Steps:

1. Read the current metadata (see Read First). Classify each opportunity as Visibility or Conversion.
2. Naming + keywords (references/aso-keywords.md): choose title/subtitle positioning; build the keyword field by relevance first, then volume, then competition; dedupe against title/subtitle/category; singular forms; no spaces after commas. Reject high-volume-but-weakly-related keywords.
3. Description + promo text (references/aso-keywords.md): hook-features-proof-CTA structure; keep the description conversion-focused, not keyword-stuffed; keep promotional text current.
4. Screenshots + previews (references/aso-product-pages.md): benefit-oriented captions; strongest screens in the first 3 slots; never lead with onboarding/splash.
5. Custom Product Pages (references/aso-product-pages.md): map each of up to 70 pages to a distinct audience/channel with tailored assets and assigned keywords; state both the 70-page capacity and assigned-keyword search visibility.
6. In-App Events (references/aso-product-pages.md): pick the badge matching real time-bound content; write name/short-description as visible search-result copy.
7. Product Page Optimization (references/aso-product-pages.md): one hypothesis, up to three treatments, selected localizations, target metric, decision rule; 90%+ confidence to call a winner.
8. Ratings + review prompts (references/aso-product-pages.md): trigger requestReview() only after a positive moment (never first launch/errors); StoreKit decides whether a prompt shows.
9. Localization (references/aso-product-pages.md): research keywords and rewrite copy per market -- never machine-translate ASO.
10. Cross-check every metadata surface against PREPARE compliance before publishing.

Output format

An ASO plan split into Visibility and Conversion, with concrete copy/keyword choices, one PPO hypothesis (up to three treatments + localizations + target metric + decision rule), a Custom Product Page-to-channel map, and a review-prompt timing note.

Mode: PREPARE (Submission Review-Prep)

Catch rejection risks before submission. Surface blockers before ordinary cleanup.

Steps:

1. Read the app target, Info.plist, entitlements, PrivacyInfo.xcprivacy, StoreKit usage, and metadata (see Read First).
2. Blocking checks first (escalate these before cleanup):

• Uploads after April 28, 2026 must be built with Xcode 26+ and the relevant platform SDK 26+.
• Privacy manifest, privacy label, privacy policy, ATT state, SDK transmissions, and audited runtime network behavior must all align.
• Digital goods/subscriptions must use StoreKit IAP; no external purchase paths/links/buttons/CTAs for digital goods unless current rules or an approved entitlement allow them.
• Required screenshot sets: 6.9-inch iPhone; 13-inch iPad when the app runs on iPad.
• Login-gated or non-obvious features need demo credentials, demo mode, and clear App Review notes.

1. Privacy manifest (references/review-privacy-manifest.md): declare each required-reason API category with an approved reason code in the bundle that uses it; SDKs carry their own manifests; do not invent broad reasons; re-check Apple's current docs.
2. IAP/StoreKit, ATT, entitlements, metadata compliance, EU DMA, submission workflow, appeals (references/review-checklists.md): work the relevant section; run the pre-submission checklist.
3. Report blockers separately from ordinary cleanup.

Output format

A readiness verdict listing blocking issues first (each with the guideline and the exact check to run), then non-blocking cleanup, then the relevant checklist from references/review-checklists.md.

Mode: AUDIT (Reviewer-Perspective Codebase Review)

Walk the actual source, config, and metadata like an App Store reviewer and report rejection risks and optimization opportunities. Change no code in the audit pass.

Follow references/audit-procedure.md end to end: identify the app's core; read the input inventory (Step 2); grep for the high-signal patterns (Step 3 -- required-reason APIs, ATT, entitlements/permission strings, IAP/restore, external purchase paths, account deletion/Sign in with Apple, regulated claims); flag top rejection risks first; run the compliance checklist; add optimization suggestions.

Output format

Findings ordered most-severe-first, each as file:line (or metadata field) + guideline reference + severity (P0/P1/P2/P3) + concrete fix, plus the full report structure in references/audit-procedure.md (executive summary, prioritized risk register, grouped detailed findings, reviewer-experience checklist, draft reviewer notes). Cite evidence for every finding or label it an Assumption with what to verify.

---

Point-in-Time Facts (re-verify before quoting)

These are current as of mid-2026 and change; confirm against Apple sources before relying on them.

• Xcode 26+ / platform SDK 26+ required for uploads after April 28, 2026.
• 6.9-inch iPhone screenshots are the primary set (as of May 2026); 6.5-inch only as fallback.
• Custom Product Pages capped at 70 per app; In-App Events up to 10 live / 15 approved.
• Apple's May 2026 fraud update: 9.1M+ submissions reviewed in 2025, 2M+ rejected (a point-in-time statistic; re-check before quoting).

Does Not Do

• HIG / native design review and "make it feel native" -> swiftui-craft. This skill carries only the review-facing HIG checklist (references/review-checklists.md), not design craft.
• Widget implementation -> widgetkit. Live Activities / Dynamic Island implementation -> activitykit. This skill covers only their App Review rules.
• Simulator/device testing, screenshots capture, permission grants at runtime -> ios-simulator.
• General StoreKit engineering (paywall code, subscription products, entitlement checks, transaction handling) -> Apple StoreKit docs. In-app review-prompt timing stays here; purchase implementation does not.
• In-app string localization (String Catalogs, RTL) -> Apple's localization docs (no dedicated skill in this repo). ASO per-market metadata stays here.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: dndungu
• Source: dndungu/agent-skills
• License: MIT

Install and usage instructions live in the source repository linked above.