The Pneumetron
NEWSTOOLSINFRASTRUCTURE
Pneumetron.Open-source guard scans iOS and Android projects before app store submission
Share
Skip to article content
  1. Home
  2. ›
  3. News
  4. ›
  5. AI Research
  6. ›
  7. Open-source guard scans iOS and Android projects before app store submission
AI Research·November 22, 2026

Open-source guard scans iOS and Android projects before app store submission

BY PNEUMETRON|4 MIN READ · 940 WORDS4 MIN READ
Tools
Share

In This Article

  • Local scanning mechanics
  • AI agent integration
  • Coverage and verification limits
  • Sources
Illustrative photo: network, server, system, infrastructure
Illustrative image: bsdrouin / Pixabay

A new open-source playbook and shell-based guard automates pre-submission checks for mobile apps, integrating with AI coding agents to block submissions against 117 known rejection patterns and 95 regulatory deadlines.

Key Takeaways

  • 01An open-source shell guard scans iOS and Android projects locally in seconds to flag critical rejection risks before submission.
  • 02The tool maps 117 documented rejection patterns to exact Apple and Google guidelines, providing concrete fixes for each.
  • 03Hooks install automatically into 13+ AI coding tools, including Claude Code, Codex, and Gemini CLI, to intercept submit commands.
  • 04The automated scan cannot verify developer account status or console registration, requiring manual checks for account readiness.

A new open-source project called app-store-compliance provides a shell-based pre-submission guard and playbook designed to catch iOS and Android app store rejections before they reach review. Maintained by a single developer, the repository bundles 117 documented rejection patterns mapped to exact Apple and Google guidelines, alongside a registry of 95 tracked regulatory deadlines covering the EU AI Act, DSA, COPPA, and various age verification laws. The tool runs locally, scanning project folders for known compliance failures, and integrates directly into AI coding agents to block submission commands when critical risks are detected.

Local scanning mechanics

The core component is a bash script that takes a path to an Xcode or Android project directory. It parses the project structure and outputs a ranked findings table, classifying issues as critical, high, or medium. Each finding includes the triggering file and line number where applicable, the exact guideline or policy reference, and a concrete fix. The script concludes with a single status line that dictates the next step: BLOCKED if a critical risk is present, CLEAR if the code passes, or NOT CHECKED if the guard cannot locate a valid project structure.

Scan times scale with project size. A small project typically completes in under a second, while larger codebases may take up to a minute. The script prints its runtime at the end of execution. For managed Expo apps lacking native folders, the guard falls back to reading app.json or app.config for purpose strings and privacy manifests, though running npx expo prebuild first ensures native code is covered. The script is designed to be non-destructive; it reads project files, writes scan results to a temporary folder, and leaves the original codebase untouched.

AI agent integration

The repository includes an installer script that configures hooks for 13+ AI coding tools, including Claude Code, Codex, Gemini CLI, Cursor, GitHub Copilot, Grok, Qwen Code, and Windsurf. The installer writes the guard into the tool's native hook directory and registers a skill or plugin that triggers the scan automatically when a submit command is issued. For supported tools, the hook intercepts the upload payload and halts the process if the guard exits non-zero.

The installation process is self-verifying. The script feeds a test submit command to the hook in the tool's native payload format and confirms it blocks successfully. It prints Install verified upon success. If a FAIL line appears, the installer suggests running with the doctor flag to diagnose the issue. The installer preserves existing configuration files, appending new entries without overwriting unrelated settings, and keeps a dated backup of any modified files. It handles tool-specific configuration paths carefully, such as writing to ~/.claude/settings.json for Claude Code, ~/.codex/hooks.json for Codex, or ~/.gemini/settings.json for Gemini CLI. For Gemini CLI versions prior to 0.26, it also enables tools.enableHooks to allow hook execution. Kimi Code receives a marked [[hooks]] block appended to its config file, while Factory Droid writes to ~/.factory/settings.json to avoid conflicts with existing hook directories.

The installer also deploys a shared skill file to ~/.agents/skills/, which every tool except Claude Code reads directly. Claude Code stores its skill in ~/.claude/skills/. The skill contains a structured prompt that instructs the agent to clone the repository, run the installer, verify the hook, execute the guard against the current project, and walk the user through the account readiness checklist. This ensures the agent handles the entire setup sequence without manual command entry.

Illustrative photo: circuit board, computer, chip, data processing Illustrative image: blickpixel / Pixabay

For Windows environments, the installer runs through Git Bash and registers hooks for both Bash and PowerShell, widening any legacy Bash-only entries. Developers can preview configuration changes with a --dry-run flag, check the installed version with --version, or remove the entire setup with an uninstall command that strips the hook and skill while preserving other tool configurations.

Coverage and verification limits

The playbook prioritizes the two most common modern rejection causes: missing or mismatched demo accounts with live backends, and privacy declarations that do not match runtime SDK usage. Each pattern in the repository traces back to live Apple or Google policy documents. The project includes a Python script, verify-citations.py, that cross-checks these citations against official sources to ensure the rules are real and not inferred. The tool also provides a pre-submission checklist for account readiness and a metadata audit script for store listings.

The automated guard has clear boundaries. It cannot verify whether a developer account is active, whether program agreements are accepted, or whether Play Console package names are registered—gaps that can delay a first launch for weeks before review even begins. Additionally, automatic hook installation is not universal. Tools like Kiro CLI, Cline, Amp, Goose, and Crush require manual JSON or configuration file edits to trigger the guard. Platforms such as Roo Code, Continue, Aider, and Copilot for Xcode lack pre-command hook support entirely, requiring developers to run the scan manually or configure git pre-push hooks before submission.

The regulatory deadline tracker monitors 95 specific dates across the EU AI Act, DSA, COPPA, and regional age laws, flagging upcoming compliance windows within a 90-day shipping radius. For non-developers, the repository provides a plain English checklist that isolates the first two critical verification steps, allowing product managers and QA teams to validate account and program readiness without touching the codebase. Developers can wire the guard into their build pipeline, while non-developers simply follow the checklist to navigate submission safely.

The repository is free and open source. Developers can clone the project, run the installer for their specific environment, and update the guard by pulling the latest changes and re-running the setup command. The project explicitly requests no payment or enterprise licensing, relying instead on community stars and forks to sustain maintenance.

Sources

  • github.com: GitHub - mjmirza/app-store-compliance: Avoid common App Store and Google Play rejections. Open playbook and pr
#appstore#googleplay#aicoding#devops#compliance
PE
PUBLISHED BY

PNEUMETRON EDITORIAL

Pneumetron's pipeline pairs AI-assisted drafting with human editorial review before publishing — our goal is to make staying informed easier, not to replace real reporting.

Source Material:github.com ↗
Share this article
Share
Stay Informed

Never miss a signal.

Subscribe to the Pneumetron Intelligence Digest.

← Previous
OpenAI blocks reasoning extraction and peer review homogenization rises
Next →
AtomicChat Releases Uncensored LoRA for Inkling-2-Mini

More from AI Research

View All →
AI Research5 min ago
Illustrative photo: laptop, macbook, codes, coding

Lang-SVG introduces hierarchical trees for semantic image vectorization

Researchers have published Lang-SVG, a method that reconstructs raster images as hierarchical, semantically labeled SVG trees rather than flat layers. The approach uses open-vocabulary segmentation and MLLMs to organize vector primitives into parent-child structures, improving both structural alignment and rendering fidelity under strict control-point budgets.

BY PNEUMETRON5 MIN READ
Read more
AI Research5 min ago
Illustrative photo: server, cloud, development, business

AtomicChat Releases Uncensored LoRA for Inkling-2-Mini

AtomicChat published a rank-1 LoRA adapter that strips refusal behavior from the 26.4B MoE Inkling-2-Mini model. The edit drops harmful refusal rates to zero while costing 1.5 points on MMLU and spiking first-token KL on previously refused prompts.

BY PNEUMETRON4 MIN READ
Read more
AI Research5 min ago
Illustrative photo: switch, it, server, server cabinet

OpenAI blocks reasoning extraction and peer review homogenization rises

OpenAI reports a 54-point drop in adversarial distillation success rates after updating GPT-6.2 safeguards, while an unreviewed arXiv study finds LLM-assisted peer reviews are converging on identical critiques.

BY PNEUMETRON4 MIN READ
Read more
AI Research5 min ago
Illustrative photo: server, technology, computer, computer technology

DaisyRose-27B RL Training, NVIDIA Dynamo, and Run:ai GPU Sharing on EKS

RadixArk releases an open-weight 27B model trained with RL, NVIDIA announces a serving layer to cut rollout bottlenecks, and Run:ai enables finer-grained GPU sharing on Amazon EKS.

BY PNEUMETRON4 MIN READ
Read more

Comments

In This Article

  • Local scanning mechanics
  • AI agent integration
  • Coverage and verification limits
  • Sources

Most Read

01
AI Research·3d ago
Tencent releases LoopForge to orchestrate coding agents in self-hosted pipelines
02
AI Research·4d ago
Recent AI Research Reveals Gaps in Coding Agents and Advances Scientific Modeling
03
AI Research·4d ago
Four New AI Papers Advance Reasoning, TTS, and Decoding Efficiency
04
AI Research·4d ago
Anthropic Upgrades Fable and Mythos as Frontier AI Expands Open-Source and Robotics
05
AI Research·3d ago
Game2Player Scales Visual Agent Training with Co-Developed Game Environments
Daily Digest

Get top AI & tech signals delivered to your inbox every morning.

Subscribe →
Follow Signals
X / TWITTERLINKEDININSTAGRAMYOUTUBETELEGRAM

THE PNEUMETRON

The publication of modern technology, AI, and infrastructure.

AI-assisted, human-reviewed.

NEWS CATEGORIES

  • AI RESEARCH
  • TECHNOLOGY
  • WORLD
  • BUSINESS
  • POLITICS
  • SCIENCE
  • HEALTH
  • SPORTS
  • ENTERTAINMENT

AI & RESEARCH

  • AI RESEARCH
  • INFRASTRUCTURE
  • TOOLS
  • DEEP DIVES
  • TECHNICAL DISPATCHES

RESOURCES

  • ABOUT US
  • CONTACT
  • CAREERS
  • ADVERTISE
Pneumetron.
© 2026 Pneumetron. AI-assisted, human-reviewed.
All rights reserved.
ABOUTTOOLSPRIVACYTERMSCONTACT