📦mjmirza / app-store-compliance
← 返回首页
Avoid common App Store and Google Play rejections. Open playbook and pre-submission guard with 117 documented rejection patterns and fixes, privacy manifest and Data safety checks, 95 tracked deadlines (EU AI Act, DSA, COPPA, age laws), and a skill for AI coding agents.
⚖️ NOASSERTION#android#app-store-review#google-play#ios#app-review-guidelines🌐 mjmirza.github.io📅 创建于 2026-06-06🔄 更新于 2026-11-19
🔗 GitHub Shell
本站解读

上架被拒往往不是功能问题,而是材料没对齐审核方的检查项。这套 playbook 把 App Store 与 Google Play 的常见拒因整理成可检索的站点,同时附带提交前的自动化 guard,以 Shell 脚本实现,可以放进提交流水线跑一遍。文档收录 117 条已记录的被拒模式及对应修法,隐私清单(privacy manifest)与 Google Play 的 Data safety 表单各有独立核对项;另有 95 项监管截止日期跟踪,涉及 EU AI Act、DSA 等。入口按使用者处境分流:即将提交、已被拒、首次上架,各自对应不同的第一步,每条路径还标注了完成第一步所需的时间。适合独立开发者、出海小团队在打包前逐项过一遍。

上榜 3 天 · 首次上榜 2026-11-19 · 近 7 天 +1.4K star · 榜单最高 #2

📄 仓库 README

在 GitHub 查看 ↗
来自 GitHub 仓库自述文件 · 版权归原作者所有
Apple   Android # App Store Compliance Playbook Stop getting your iOS and Android apps rejected. The enterprise reference and automated guard that turns App Store and Google Play rejection into a designed out failure mode. Read it as a website at [mjmirza.github.io/app-store-compliance](https://mjmirza.github.io/app-store-compliance/). [![Apple App Store](https://img.shields.io/badge/Apple_App_Store-000000?logo=apple&logoColor=white)](https://developer.apple.com/app-store/review/guidelines/) [![Google Play](https://img.shields.io/badge/Google_Play-414141?logo=googleplay&logoColor=white)](https://play.google/developer-content-policy/) [![Claude Code](https://img.shields.io/badge/Claude_Code-ready-000000?logo=anthropic&logoColor=white)](#how-you-actually-use-this) [![OpenRoots ORA 2.3](https://openroots.org/badge/ora.svg)](https://openroots.org/licenses/ora/2.3) [![CI](https://img.shields.io/github/actions/workflow/status/mjmirza/app-store-compliance/ci.yml?branch=main&label=CI)](https://github.com/mjmirza/app-store-compliance/actions) [![PRs welcome](https://img.shields.io/badge/PRs-welcome-2ea44f)](https://github.com/mjmirza/app-store-compliance/blob/HEAD/.github/CONTRIBUTING.md) [![Stars](https://img.shields.io/github/stars/mjmirza/app-store-compliance?style=social)](https://github.com/mjmirza/app-store-compliance/stargazers) [![Forks](https://img.shields.io/github/forks/mjmirza/app-store-compliance?style=social)](https://github.com/mjmirza/app-store-compliance/fork) [![Follow mjmirza](https://img.shields.io/github/followers/mjmirza?label=Follow&style=social)](https://github.com/mjmirza) **If this saves you even one rejection, leave a star, fork it, and follow along. That is the whole ask.** [Star this repo](https://github.com/mjmirza/app-store-compliance)  |  [Fork it](https://github.com/mjmirza/app-store-compliance/fork)  |  [Follow on GitHub](https://github.com/mjmirza)  |  [Follow on X](https://twitter.com/MirzaJhanzaib)  |  [next8n.com](https://next8n.com)

Start here

Pick one row. The time shown is for that first step.

You are Do this first Time
About to submit, and you can open a terminal Run the two commands below 2 minutes
Already rejected Open the appeal playbook 10 minutes
Not a developer Open the checklist and do its first two checks 5 minutes
Using an AI coding tool (Claude Code, Codex, Gemini CLI, GitHub Copilot, Cursor, Grok, Qwen, Kimi and more), and want this on every submit Go to the install 1 minute

The two commands. Needs git and bash. The first downloads this playbook. The second scans your app. Nothing is installed and nothing in your app is changed.

git clone https://github.com/mjmirza/app-store-compliance ~/repositories/app-store-compliance
bash ~/repositories/app-store-compliance/agent-os/hooks/app-store-compliance-guard.sh "/path/to/your/app"

Replace only /path/to/your/app, with the folder that holds your Xcode project or your Android app. Keep the quotes, they keep a folder name with a space in it whole. To get the path without typing it, drag the folder from Finder or your file manager into the terminal window. A small project scans in about a second, a large one can take up to a minute, and the guard prints the time it took. Findings print first, critical ones on top, each with the file and line that triggered it when there is one. The last lines tell you where you stand.

Last line starts with It means Do next
BLOCKED. A critical rejection risk was found Fix each [CRITICAL] line using the fix. line under it, then run the same command again
CLEAR. No critical risk in your code Do the account and listing checks in the checklist
NOT CHECKED. No iOS or Android project was found in that folder, or the guard could not write its scan files. Run by hand this exits 1. As a hook it lets the command through Point it at the right folder, and check the temp folder is writable. A managed Expo app with no native folders still gets its app.json or app.config read for purpose strings, the privacy manifest and Android permissions. Run npx expo prebuild to cover the native code too

What you get

A rejection costs a week. You submit, you wait days in review, you get a one line rejection, you guess at the fix, you resubmit, you wait again. This playbook turns that loop into a check you run before you ever submit.

  • You catch the rejection before Apple or Google does. The guard runs against your project and names the risk while you can still fix it cheaply, not after a reviewer has already said no.
  • You stop guessing at the cause. Every rejection pattern maps to the exact guideline and the exact fix, so a one line rejection email becomes a known problem with a known answer.
  • The two checks that cause most rejections are covered first. A working demo account with a live backend, and a privacy declaration that matches what your app actually does at runtime, including every SDK. These are the top modern rejection causes and the playbook holds them above everything else.
  • It works whether or not you write code. A developer wires the guard into the build. A non developer reads the plain English checklist and walks the submission safely.
  • Every claim is real. Guideline numbers and statistics trace to a live Apple or Google source, never invented, so you are acting on the actual rules, not a guess.

If it saves you one rejection, it has paid for the ten minutes it takes to read.

Install it into your AI coding tool

This puts the check in front of every submit. Your agent runs the guard before any submit command, such as an upload to App Store Connect or Google Play, and stops the command when a critical rejection risk stands. For a one-time scan, skip this section and use Start here.

One command per tool, about a minute. Needs git and bash. Every tool other than Claude Code also needs python3.

git clone https://github.com/mjmirza/app-store-compliance ~/repositories/app-store-compliance
bash ~/repositories/app-store-compliance/scripts/install.sh --tool all

--tool all installs into every supported tool it finds in your home folder. To pick one, pass its name instead.

Tool Command What it gets One thing to do after
Claude Code install.sh Hook, skill, /app-store-audit Restart Claude Code
Codex CLI install.sh --tool codex Hook, skill Run /hooks in Codex once and approve the new hook. Codex runs no new hook until you do
Gemini CLI install.sh --tool gemini Hook, skill, /app-store-audit Restart Gemini CLI. The hook works from 0.22, the skill needs 0.24 or later
OpenCode install.sh --tool opencode Plugin, skill, /app-store-audit Restart OpenCode
Cursor install.sh --tool cursor Hook, skill Restart Cursor
GitHub Copilot CLI and Copilot in VS Code install.sh --tool copilot Hook, skill Start a new Copilot CLI session and reload VS Code
GitHub Copilot cloud agent and Copilot in JetBrains install.sh --tool copilot-repo --repo "/path/to/your/app" Hook in your app's .github/hooks Commit .github/hooks to your default branch. The cloud agent reads hooks only from there
Grok install.sh --tool grok Hook, skill Start a new Grok session
Qwen Code install.sh --tool qwen Hook, skill, in $QWEN_HOME when it is set Restart Qwen Code
Kimi Code install.sh --tool kimi Hook, skill Restart Kimi Code
Factory Droid install.sh --tool droid Hook, skill Start a new Droid session
Augment Auggie install.sh --tool auggie Hook, skill Start a new Auggie session
Windsurf install.sh --tool windsurf Hook, skill Restart Windsurf

Every run ends with Install verified. only after it has fed a submit command to the hook in that tool's own payload format and seen it blocked. For OpenCode that check runs the plugin under node, so without node the run says Installed, but not fully verified. instead. If you see a FAIL line instead, fix that line and run the same command with doctor added.

What it writes, so you can check it.

  • The guard, into that tool's own folder, for example ~/.codex/hooks/ or ~/.gemini/hooks/.
  • The skill. Claude Code reads it from ~/.claude/skills/. Every other tool reads one shared copy in ~/.agents/skills/.
  • One hook entry in that tool's config (~/.claude/settings.json, ~/.codex/hooks.json, ~/.gemini/settings.json, ~/.cursor/hooks.json, ~/.copilot/hooks/ or $COPILOT_HOME/hooks/), or for OpenCode one plugin file in ~/.config/opencode/plugins/. Every other setting is kept, and a dated backup sits next to a file that already existed. A config that is not valid JSON or repeats a key is never rewritten. You get the exact block to paste instead.
  • For Gemini CLI it also turns on tools.enableHooks, which Gemini needs before version 0.26 to run any hook.
  • For Kimi Code it appends one marked [[hooks]] block to ~/.kimi-code/config.toml (or $KIMI_CODE_HOME/config.toml) and never rewrites or prints the rest of that file, which can hold API keys. Uninstall removes exactly that block.
  • For Factory Droid it writes to ~/.factory/settings.json when your hooks already live there, because Droid ignores them once ~/.factory/hooks.json exists.

Tools the installer does not wire, and what to do instead

Each of these was checked against its own documentation. Where a tool can block a command but its format could not be confirmed, you get the manual step rather than a guess.

Tool What to do
Kiro CLI Kiro sends the event as JSON on stdin and blocks the tool when a PreToolUse command hook exits non-zero, and the guard reads Kiro's shell, execute_bash, execute_pwsh and execute_cmd payloads. The installer does not wire it yet. After install.sh, add .kiro/hooks/app-store-compliance.json to the project with {"version":"v1","hooks":[{"name":"app-store-compliance","trigger":"PreToolUse","matcher":"^(shell\|execute_bash\|execute_pwsh\|execute_cmd)$","action":{"type":"command","command":"bash ~/.claude/hooks/app-store-compliance-guard.sh"}}]}. Not yet tested agains

📈 近 7 天 Star 趋势

每日新增 Star(柱)与累计 Star(线)· 数据来自 GitHub

🏆 排名变化趋势

在各榜单中的排名历史(数字越小越靠前)

📋 历史上榜记录

每次快照的详细数据
日期 榜单 排名 周期新增 总 STAR