App Store Compliance: Guard iOS and Google Play releases with 117 rejection patterns
A preflight checker for mobile teams targeting App Store and Google Play, combining code scans, deadlines, and AI disclosure checks.
GitHub mjmirza/app-store-compliance Updated 2026-11-20 Branch main Stars 1.8K Forks 126
Shell/Python App Store and Google Play compliance 117 rejection patterns AI coding agents and mobile apps

🧭 Decision Guide

Try it if you

  • You build iOS and Android apps with Flutter, React Native, or Expo.
    README's docs/CROSS-PLATFORM-FRAMEWORKS.md lists these frameworks and says the guard scans the built artifact surface.
  • Your app calls external AI or handles personal data and you need to check disclosure of the AI provider.
    The README's Why this is urgent right now (2026) section says that since late 2025, missing a consent modal naming the provider leads to rejection.
  • Your team needs to scan multiple apps and retain per-occurrence waivers plus a GitHub Code Scanning export.
    README's docs/FLEET-SWEEP.md describes Fleet Sweep, per-occurrence waivers, and GitHub code-scanning export.
  • You need to track effective and enforcement dates for the EU AI Act, DSA, or COPPA.
    README lists data/regulatory-deadlines.json and docs/EU-REGULATORY-2026.md and docs/GLOBAL-REGULATORY-2026.md.

Skip it if you

  • Your target is not Apple or Google Play and you publish only to stores outside the documented coverage.
    The README focuses on App Store and Google Play; other stores appear as supplementary coverage in docs/OTHER-STORES.md.
  • You need fully automatic decisions for store listings, account status, and other non-code requirements.
    README's data/detection-recipes.json says 107 of 114 items are code-detectable and the other seven are manual store listing and account checks.
  • You need a complete live regulatory-announcement feed rather than simulated migration output.
    README says PRIVACY-POLICY-MIGRATION.md and SECURITY-POLICY-MIGRATION.md contain simulated output, not live announcements.

Requirements

  • The environment must run Shell or Python; the project lists Shell and Python as its technologies.
  • For AI coding agent integration, use agent-os/skill/SKILL.md or agent-os/hooks/app-store-compliance-guard.sh.
  • Deadline checks depend on data/regulatory-deadlines.json; the README does not specify a Python version.
  • The 2.4.0 release notes mention an iOS 13 floor, IPv6, debuggable, and 64-bit ABI checks.

First step (verbatim from README)

scripts/monitor-privacy.py --simulate

Watch out

  • Do not treat monitor-privacy.py --simulate as live privacy-announcement synchronization.
    The README explicitly says the migration output is simulated and not live announcements.
  • Do not treat code-scan results as a complete approval verdict; seven store listing and account checks remain manual.
    README's data/detection-recipes.json says the other seven items cannot be completed by code scanning.
  • AGENTS.md says monitor corroboration is keyword-based, so cited official sources still require human reading.
    The README's description of AGENTS.md includes this verification rule.
  • The 117 documented rejection patterns and 114 items in detection-recipes.json use different scopes and should not be treated as the same count.
    The project description says 117 documented rejection patterns, while the README says 107 of 114 are code-detectable.

Alternatives

  • fastlane precheck:It is lighter when you only need App Store metadata rules rather than 117 rejection patterns or a regulatory deadline database.
    README's docs/OPEN-SOURCE-PATTERNS.md
  • Google Play pre-launch report:It is more direct when the priority is runtime issues found by Google Play test devices rather than cross-platform compliance rules.
    README's docs/OPEN-SOURCE-PATTERNS.md
  • Apple App Store Review Guidelines and Google Play policy center:They are better when you need to read the platforms' primary policies directly rather than use the project's curated rejection maps.
    README's Sources section

Not stated in the README

  • The excerpt does not provide complete installation commands, supported Shell/Python versions, or an operating-system matrix.
  • The excerpt does not specify the full parameters, exit codes, or CI integration examples for app-store-compliance-guard.sh.
  • The excerpt does not quantify scan time, false-positive rate, false-negative rate, or resource usage for large apps.
  • The excerpt does not provide the exact text of the Other license or its commercial-use restrictions.
  • The excerpt does not show what proportion of Apple and Google Play rules the 117 rejection patterns cover.
  • The excerpt does not describe compatibility changes or rollback procedures after 2.4.0.

💡 Deep Analysis

6
It depends My AI app sends personal data to an external AI provider, was developed with Claude Code, and targets EU users. Can I use this project to handle AI disclosure, privacy manifests, and EU AI Act deadlines together?
For: An owner of an AI mobile app that sends personal data to an external AI provider, targets the EU, and uses Claude Code for development

It depends. The project can consolidate these risks into a pre-submission checklist, but it cannot establish your actual data facts or provide a legal conclusion.

  • The README states that since late 2025, sending personal data to an external AI requires a consent modal naming the provider, otherwise the app may be rejected.
  • docs/ADVANCED-2026.md covers Privacy Manifest, AI content policy, and the legal layer; docs/EU-REGULATORY-2026.md lists the EU AI Act Article 50 transparency deadline of 2 August 2026.
  • The project also provides AI policy migration monitoring and a regulatory-deadline database for tracking policy changes and effective dates.
  • However, static scanning cannot establish whether personal data is actually sent, whether consent is valid, or how the model provider processes it; monitor-privacy.py --simulate is explicitly simulated output.
  • README, “Why this is urgent right now (2026)”: an app sending personal data to an external AI without a consent modal naming the provider is rejected
  • README, “What is inside”: `docs/EU-REGULATORY-2026.md` lists EU AI Act Article 50 transparency by 2 August 2026
  • README, “What is inside”: `docs/PRIVACY-POLICY-MIGRATION.md` is generated by `scripts/monitor-privacy.py --simulate`, not live announcements
  • Project insights, “usage_limitations”: monitoring scripts and regulatory-date databases require ongoing source verification
scripts/monitor-privacy.py --simulate
Not stated in the README:The README does not state how actual data flows are used to identify external AI providers and personal-data categories automatically;It does not state whether a consent-modal template, detection rule, or platform-specific implementation is provided;The project does not replace EU privacy or AI legal advice
Yes I manage multiple iOS and Android apps and want Fleet Sweep to produce one unified check and export results to GitHub Code Scanning. Is this project suitable as a multi-app release gate?
For: An enterprise mobile-platform owner responsible for releasing multiple iOS and Android apps who plans to use Fleet Sweep for one unified release verdict

Yes, but it should be treated as a multi-app release risk gate, not as proof of store approval.

  • README docs/FLEET-SWEEP.md explicitly supports scanning “every app you ship” and producing one release verdict.
  • Fleet Sweep supports expiring, per-occurrence waivers for reviewed false positives, which is more controlled than disabling an entire rule category permanently.
  • The README also lists a GitHub code-scanning export, allowing integration with an existing enterprise security workflow.
  • Rule data, citations, regulatory dates, and monitoring scripts can change; the project analysis also says enterprises must verify the terms of the Other license and the complete LICENSE file before adoption.
  • README, “What is inside”: `docs/FLEET-SWEEP.md` supports one release verdict across every app you ship
  • README, “What is inside”: Fleet Sweep includes expiring per-occurrence waivers and a GitHub code-scanning export
  • Project insights, “architectural_strengths”: rules and execution are separated, with auditability and traceability
  • Project data: license is `Other`
Not stated in the README:The excerpt does not provide the complete Fleet Sweep command, input directory structure, or concurrency model;The exact SARIF version and permissions required for GitHub Code Scanning export are not stated;The scale at which scan time and resource consumption become limiting is not stated
Yes I am releasing a kids' or health app to both the App Store and Google Play. Can this project cover COPPA, age ratings, privacy, and platform policies rather than merely scanning permissions?
For: A mobile product owner releasing a kids' or health app who must handle COPPA, age ratings, privacy, and cross-platform store rules

Yes, for a pre-submission risk checklist covering these topics, but a kids’ or health app cannot be reduced to permission scanning.

  • README docs/BY-APP-TYPE.md routes checks by app type and explicitly includes kids, health, subscriptions, social, AI, finance, and VPN apps.
  • docs/GLOBAL-REGULATORY-2026.md covers US COPPA, state app-store age laws, and other jurisdictions; the Apple and Google rejection maps are in docs/APPLE.md and docs/GOOGLE-PLAY.md.
  • docs/PRE-SUBMISSION-CHECKLIST.md provides exhaustive pre-submission checks for both stores, while docs/ADVANCED-2026.md covers age ratings and privacy requirements.
  • Actual user ages, health-data purposes, parental consent, moderation, and privacy-policy accuracy still require human evidence; code scanning alone cannot confirm them.
  • README, “What is inside”: `docs/BY-APP-TYPE.md` covers kids and health applications
  • README, “What is inside”: `docs/GLOBAL-REGULATORY-2026.md` covers US COPPA and state app-store age laws
  • README, “What is inside”: `docs/PRE-SUBMISSION-CHECKLIST.md` provides exhaustive checklists for both stores
  • Project insights, “usage_limitations”: code scanning cannot reliably determine business intent, valid user consent, or whether moderation actually operates
Not stated in the README:The README does not state the automation ratio for kids- or health-app rules;It does not state whether App Store Connect and Play Console age ratings, privacy questionnaires, and store configuration can be read automatically;Health-data and child-consent requirements in individual countries may exceed the documented scope
Yes I maintain Flutter, React Native, and Expo apps and produce both IPAs and APKs/AABs. Does this project inspect the final build artifacts, or does it only inspect framework source code?
For: A cross-platform mobile engineer maintaining Flutter, React Native, or Expo apps and producing both iOS IPAs and Android APKs/AABs

Yes. It is suitable because it explicitly targets the final build artifact surface rather than only analyzing Flutter, React Native, or Expo source code.

  • The README lists Flutter, React Native, Expo, Capacitor, Ionic, Cordova, Unity, .NET MAUI, Tauri mobile, and Kotlin Multiplatform.
  • The project analysis states that permissions, APIs, resources, and configuration in the uploaded package are more representative of store-review risk than framework-level source alone.
  • This does not provide complete release validation: store descriptions, screenshots, privacy-policy accuracy, account configuration, and actual runtime data processing still require separate verification.
  • The excerpt does not provide the complete commands or exact entry points for scanning APK/AAB, IPA, Info.plist, AndroidManifest, and dependency SDKs.
  • README, “What is inside”: `docs/CROSS-PLATFORM-FRAMEWORKS.md` covers Flutter, React Native, Expo, and other frameworks
  • Project insights, “architectural_strengths”: permissions, APIs, resources, and configuration in the uploaded package better reflect review risk
  • README, `RULING-REGISTER.md`: Release requirements a code scan cannot see
  • Project insights, “usage_limitations”: visibility into closed-source SDKs, remote configuration, and runtime behavior is limited
Not stated in the README:The excerpt does not specify per-framework input files, artifact parameters, or scan commands;It does not state whether encrypted, obfuscated, or specially produced artifacts can be fully unpacked and analyzed;Coverage of third-party SDK risk rules for the target versions is not stated
It depends I used Claude Code to rapidly build an app for both iOS and Android, but I have no dedicated compliance specialist. Can this project find the major rejection risks before I submit to the App Store and Google Play?
For: An independent developer using Claude Code to build an iOS and Android app in an afternoon without a dedicated compliance specialist

It depends: the project is suitable for pre-submission risk screening, but it cannot replace store review or manual verification.

  • It covers Apple App Store and Google Play and documents 117 rejection patterns with fixes.
  • The README includes checks for Privacy Manifest, Google Play Data Safety, permissions, export compliance, payments, age ratings, and AI content disclosure.
  • It is designed to inspect the final upload artifact rather than only the framework source produced by Claude Code, so generated permissions, SDKs, and build configuration matter.
  • The truthfulness of the privacy policy, store metadata, account status, actual data flows, and moderation process cannot be reliably established by code scanning alone.
  • Project description: 117 documented rejection patterns and fixes
  • README, “What is inside”: `docs/ADVANCED-2026.md` covers privacy manifests, export compliance, payments, legal requirements, and AI content policy
  • Project insights, “architectural_strengths”: scans the built artifact surface rather than only framework source
  • README, `RULING-REGISTER.md`: Release requirements a code scan cannot see
Not stated in the README:The supplied README excerpt does not include the complete installation command or the exact Claude Code skill installation procedure;Coverage for the specific third-party SDKs used by the target app is not stated;False-positive and false-negative rates for an individual developer environment are not stated
Yes I maintain a CI/CD environment primarily using Shell and Python and want to integrate App Store and Google Play checks into GitHub Actions and an AI coding-agent hook. Does this project fit that integration model?
For: An enterprise CI/CD engineer who wants to integrate mobile compliance into GitHub Actions and AI coding-agent hooks, primarily maintaining Shell and Python tooling

Yes. Its script-oriented structure and agent integration model fit this CI/CD constraint, although environment compatibility must be verified.

  • The project’s main language is Shell, with substantial Python content; the README lists multiple scripts/monitor-*.py monitoring scripts.
  • The project analysis states that the tools can run independently or integrate with CI/CD, GitHub Actions, pre-commit hooks, and AI coding agents.
  • The README describes an AI agent skill, Fleet Sweep, and GitHub Code Scanning export, covering different entry points from code changes to multi-app release verdicts.
  • However, the project depends on Shell, Python, mobile build tools, and platform CLI environments; compatibility with Windows, restricted CI runners, or unusual mobile build chains is not specified in the excerpt.
  • Project data: main language is Shell; language distribution includes Shell and Python
  • README, “What is inside”: `scripts/monitor-ai-policy.py`, `scripts/monitor-android.py`, and `scripts/monitor-privacy.py --simulate`
  • Project insights, “architectural_strengths”: Shell and Python tools can run independently or integrate with agent hooks, CI/CD, or GitHub Code Scanning
  • Project insights, “usage_limitations”: depends on Shell, Python, build tools, and platform CLI environments
scripts/monitor-privacy.py --simulate
Not stated in the README:The excerpt does not provide complete configuration files for the AI-agent skill, hooks, or GitHub Actions;Required Python, Shell, macOS, or Android SDK versions are not stated;Exit codes, report formats, and blocking behavior for CI failures are not stated

✨ Highlights

  • Documents 117 rejection patterns across Apple and Google Play
  • Checks Privacy Manifest, Data safety, and AI disclosure
  • Tracks 95 regulatory deadlines, including the EU AI Act and DSA
  • Fleet Sweep scans multiple apps before a release

🔧 Engineering

  • Runs the pre-submission compliance guard with app-store-compliance-guard.sh
  • Drives detectable rejection rules from rejection-patterns.json
  • Scans built artifacts for Flutter, React Native, Expo, and other frameworks
  • Reads global regulatory dates with deadline-checker.py

⚠️ Risks

  • The README says apps sending personal data to external AI need provider disclosure or face rejection since late 2025
  • The README identifies missing Privacy Manifest as a major modern Apple upload rejection
  • Code scanning covers 107 of 114 items; seven still require manual checks
  • The Other license and only 2 contributors limit visible governance information

👥 For who?

  • Teams shipping mobile apps with Flutter, React Native, or Expo
  • Developers rapidly shipping iOS apps with Claude Code or other AI coding tools
  • Teams maintaining multiple apps and needing Fleet Sweep release verdicts
  • Compliance owners tracking EU AI Act, COPPA, or DSA deadlines