store-compliance-pro
Skill5 starsChanged 11 days agoMITView on GitHub
Audits an Apple or Android app and its store listings for region-specific rejection risks — third-party AI brand mentions, generative-AI and licensing rules (China mainland ICP and CAC filings, game licenses, gambling, loans, VPNs), EU trader status, storefront-based feature gating, privacy consent order for Chinese Android stores, and metadata per storefront — and drives the fix after a regional App Store or Play rejection. Use before shipping to a new region, when adding an AI provider or regulated feature, or when a submission is rejected for one country or storefront.
In this repository19 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- app-intents-proSkill
- coredata-swift6-proSkill
- instruments-profiling-proSkill
- motion-design-proSkill
- swift-accessibility-proSkill
- swift-adaptive-layout-proSkill
- swift-concurrency-proSkill
- swiftdata-proSkill
- swift-design-system-proSkill
- swift-docc-proSkill
- swift-error-handling-proSkill
- swift-logging-proSkill
- swift-memory-storage-proSkill
- swift-modularization-proSkill
- swift-package-proSkill
- swift-testing-proSkill
- swiftui-proSkill
- xcode-security-proSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 15
Audit: before a release, before adding a storefront or country, or after adding an AI provider, payments, or another regulated feature.
- Very specific: line 17
Set up: no register exists yet. Create docs/compliance/regions.md and the RegionPolicy seam per references/rejection-response.md ("Register …
- Distinctive: line 19
If doing a partial task, load only the relevant reference file.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Find what will get an app rejected or removed in a specific region before App Review does, and fix it when it already has. This is the deep companion to the always-on apple-regional-compliance.md and android-regional-compliance.md rules. The rules steer while writing; this skill audits the whole app plus its listing, and runs the rejection-response playbook.
Pick the mode from the request:
- Audit: before a release, before adding a storefront or country, or after adding an AI provider, payments, or another regulated feature. Run
references/audit.mdand check each in-scope region againstreferences/regions.md. - Respond: a submission was rejected or removed for a region. Follow
references/rejection-response.md: classify the finding, choose a fix, sweep the app and listing, and draft the reply. - Set up: no register exists yet. Create
docs/compliance/regions.mdand theRegionPolicyseam perreferences/rejection-response.md("Register and policy seam").
If doing a partial task, load only the relevant reference file.
Core Instructions
- Region is decided per storefront, and the store decides. Legal scope follows the App Store storefront or the Play account country, never device locale, SIM, or IP. Gating on locale is a finding.
- The listing is part of the submission. Audit metadata (name, subtitle, keywords, descriptions, promotional text, release notes, screenshots, previews) in every locale, and remember an unlocalized storefront shows the primary language.
- Third-party AI brands are the top regional trigger. ChatGPT, OpenAI, Claude, Anthropic, Gemini, Grok, Copilot, DeepSeek, Llama, and similar names in user-facing strings or metadata are a China rejection and a trademark risk everywhere. Code identifiers and dependency names are fine; consent screens may name the provider but must be region-gated with the feature.
- One policy seam. Regulated features ask a single injected
RegionPolicy. Scattered country-code comparisons, or none at all, are findings. - Prefer the fix that removes the risk. Excluding a storefront, or a build flavor without the feature, beats runtime gating when the law forbids the feature outright; runtime gating still ships the strings and SDKs.
- Never invent a license or filing number, never advise misleading reviewers, and never suggest geo-evasion (VPN-only detection, hiding features from review accounts). Compliance claims in review notes must be true.
- Mark uncertainty. Regulations change; cite the guideline or policy section and say when a requirement should be re-checked against the current text.
- Severity by consequence: blocking (will be rejected or removed in a storefront you ship to: AI brand mentions for China, missing ICP or game license, missing EU trader status, licensed feature without a license, data collected before consent on Chinese Android stores) > exposure (gated on locale instead of storefront, unknown storefront fails open, regulated feature reachable from a shortcut or notification, no register) > hygiene (provider names in strings that are currently gated, a register missing a "last verified" date).
- Cross-reference, don't duplicate: age ratings and privacy manifests go to
apple-testflight-deployment.md; Data safety toandroid-play-beta-deployment.md; AI consent and provider wiring toapple-foundation-models.mdandandroid-ai-best-practices.md; generic strings to the localization rules.
Output Format
Audit mode: a region table first (storefront or country · shipping? · requirements · status), then findings grouped by severity. Each finding gives the location (file:line, or the metadata field and locale), the region and the rule or guideline it breaks, why review will catch it, and the fix with before/after. End with the scan counts and the first three fixes.
Respond mode: the rejection classified (region, guideline, trigger), the chosen fix and why, every location changed, the review-notes text, the Resolution Center or Play appeal reply, and the register entry.
Example finding:
Blocking — China mainland: third-party AI brand in UI and metadata
Localizable.xcstrings key chat.title reads "Ask ChatGPT", and the English description says "Powered by ChatGPT". The China storefront has no Simplified Chinese localization, so App Review there reads the English text. Generative-AI services need a CAC filing in China and ChatGPT has none (Guideline 5).
// Before
static let title = LocalizedStringResource("chat.title", defaultValue: "Ask ChatGPT")
// After — generic label; the provider is named only on the region-gated consent screen
static let title = LocalizedStringResource("chat.title", defaultValue: "Ask the AI assistant")Then either exclude China mainland in Pricing and Availability, or gate the feature with RegionPolicy.allows(.externalAIProviders) and say so in the review notes. Remove "Powered by ChatGPT" from every metadata locale and from screenshot 3.
End of example.
References
references/regions.md— per-region requirements for both stores: China mainland (Apple and Chinese Android stores), the EU, Japan, South Korea, age-assurance regions, and licensed categories, with where each is declared.references/audit.md— scan commands for AI brand mentions in code, strings, and metadata; storefront-gating checks; consent-before-collection checks for Android; and an App Store Connect / Play Console checklist.references/rejection-response.md— classifying a regional rejection, choosing among exclude / gate / license, the sweep, reply and review-notes templates, and the register andRegionPolicyseam.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

