Reversing React Native Apps logo

Reversing React Native Apps

Community
trilwu
reversing-react-native-apps

Reverse engineer React Native mobile apps, including Hermes bytecode bundles, using hbctool, hermes-dec, and Frida. Use when an APK contains index.android.bundle or libhermes.so, when an IPA contains main.jsbundle, when jadx shows only ReactActivity classes, or when a bundle file starts with the Hermes magic bytes instead of readable JavaScript.

Overview

Publishertrilwu
Repositorysecskills
Skill namereversing-react-native-apps
Stars
144
Forks
15
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by trilwu on GitHub. Read the source before you install it.

Installation

Install the Reversing React Native Apps AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/trilwu/secskills.git /tmp/secskills
mkdir -p .claude/skills
cp -r /tmp/secskills/secskills-core/skills/reversing-react-native-apps .claude/skills/reversing-react-native-apps
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Reversing React Native Apps in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Reversing React Native Apps on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Reversing React Native Apps is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Reversing React Native Apps

React Native apps are JavaScript wearing a native shell. That is good news — the application logic ships as a bundle you can read — unless the bundle was compiled to Hermes bytecode, in which case you need a decompiler and the version compatibility problem becomes the whole task.

When to Use

  • jadx shows ReactActivity, ReactNativeHost, and little application logic
  • The APK contains assets/index.android.bundle or lib/*/libhermes.so
  • The IPA contains main.jsbundle
  • You need to find API endpoints, secrets, or business logic in an RN app
  • You need to patch or hook JavaScript-level behaviour

When NOT to Use

  • Flutter (libapp.so, libflutter.so) — use reversing-flutter-apps
  • Unity (global-metadata.dat) — use reversing-unity-il2cpp
  • Native Java/Kotlin apps — use testing-mobile-applications
  • Native modules written in C/C++ — use analyzing-binaries

Identify the Bundle Format

This one check decides whether the job takes ten minutes or a day.

bash
unzip -j target.apk 'assets/index.android.bundle' -d ./out
file ./out/index.android.bundle
xxd ./out/index.android.bundle | head -2
First bytesFormatPath
Readable JS (var __BUNDLE_START, __d(function)Plain JavaScriptBeautify and read directly
Binary, Hermes magic (c6 1f bc 03 little-endian)Hermes bytecodeDecompile — see below
bash
# Plain bundle: beautify, then it reads like any minified web app
npx js-beautify index.android.bundle -o bundle.js
rg -n 'https?://|api[_-]?key|Bearer |secret|token' bundle.js | head -50

A plain bundle is a gift. Module boundaries survive as __d(function(...) registrations, source maps are occasionally shipped by mistake (index.android.bundle.map — always check), and the whole application logic is in front of you.

Hermes Bytecode

Hermes compiles JS to its own bytecode (HBC). The bytecode version is embedded in the header and changes with React Native releases, which is the single biggest practical obstacle: a tool that supports HBC 74 will refuse or misparse HBC 96.

bash
# Read the version out of the header before choosing a tool
hbctool disasm index.android.bundle ./disasm     # HBC 59, 62, 74 ONLY
hermes-dec ...                                    # decompiler; wider version support

Upstream hbctool ships support for exactly three bytecode versions — 59, 62, and 74. Anything else fails at parse. Check the header version first rather than reading a parse error as "the bundle is protected"; on a current React Native release you will usually be above 74 and should reach for hermes-dec or a hermesc you built from the matching Hermes tag.

ToolGives youCaveat
hbctoolDisassembly and reassembly — you can patch and repackSupports a limited set of HBC versions
hermes-decDecompiled pseudo-JavaScriptRead-only; output is approximate
hasmerDisassembly/assembly, alternate version coverageTry when hbctool refuses the version

When every tool rejects the version, the reliable fallback is to build the matching hermesc from the React Native release the app used, and use its tooling. That is slower but version-correct.

What to do with the output. Even imperfect decompilation is enough for the questions that matter: string tables survive intact, so endpoints, keys, and feature flags are recoverable directly:

bash
strings -n 6 index.android.bundle | rg -i 'https?://|api|token|secret|firebase' | sort -u

The Hermes string table is a flat, readable region of the file. Reach for it before decompiling — it often answers the question outright.

Runtime Analysis

Frida is frequently faster than static work on RN, because the interesting boundary is where JavaScript calls into native.

bash
# Enumerate the native modules the JS side can reach
frida -U -f com.target.app -l rn-enumerate-modules.js

# Hook the bridge: every JS↔native call, with arguments
# (target the ReactNative bridge / TurboModule dispatch)
objection -g com.target.app explore

Useful hook points: fetch/XMLHttpRequest in the JS runtime, the native networking module (OkHttp on Android — hookable with standard pinning bypass), AsyncStorage reads and writes, and any NativeModules.* the app defines.

TLS interception is ordinary here — unlike Flutter, RN uses the platform HTTP stack, so the usual system-CA plus proxy setup works, and standard pinning bypasses apply. If interception fails, suspect pinning in a native module, not a separate trust store.

Where the Findings Usually Are

React Native apps concentrate the same few problems:

  • Secrets in the bundle. API keys, Firebase configs, and third-party tokens shipped in JS because "it's compiled." Hermes is not encryption.
  • Client-side authorization. Role checks and feature gates implemented in JS, trivially patched or simply ignored by calling the API directly.
  • AsyncStorage as a credential store. Unencrypted by default; tokens and PII routinely land there.
  • Over-broad native modules. Custom bridge methods that expose file system or shell access to JS, reachable from any injected script.
  • Debug artifacts. Source maps, dev-mode bundles, and console.log output left in release builds.

Patching and Repacking

bash
hbctool disasm index.android.bundle ./work
# edit ./work/instruction.hasm and ./work/metadata.json
hbctool asm ./work index.android.bundle.patched
# replace in the APK, then re-sign
apktool b target -o patched.apk && apksigner sign --ks debug.keystore patched.apk

Patch only what you must, and only with authorization. Repacking trips integrity checks in hardened apps; if the app validates its own bundle, hook the check rather than defeating it by editing.

Rationalizations to Reject

  • "The bundle is binary, so the logic is protected." Hermes is a compiler, not a protection. The string table alone often gives up the API.
  • "jadx found nothing, so it's obfuscated." There is nothing in the dex. Look in assets/.
  • "No tool supports this HBC version, dead end." Build hermesc from the matching React Native release, or read the string table.
  • "The key is only used client-side." It is in the bundle, so it is public. Treat every bundle secret as disclosed.
  • "The app enforces the role check." The app is the client. Verify the check server-side by calling the API without it.

References

  • testing-mobile-applications — the wider assessment, storage, and platform issues
  • testing-apis — the backend, which is where the real findings are
  • reversing-flutter-apps — the other common cross-platform framework
  • hbctool, hermes-dec, hasmer, Frida/objection, jadx, apktool

Frequently asked questions

What does the Reversing React Native Apps AI skill do?

Reverse engineer React Native mobile apps, including Hermes bytecode bundles, using hbctool, hermes-dec, and Frida. Use when an APK contains index.android.bundle or libhermes.so, when an IPA contains main.jsbundle, when jadx shows only ReactActivity classes, or when a bundle file starts with the Hermes magic bytes instead of readable JavaScript.

Why use Reversing React Native Apps on TypingMind?

Because you install it once and use it with any model. Reversing React Native Apps is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Reversing React Native Apps in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/trilwu/secskills/tree/main/secskills-core/skills/reversing-react-native-apps. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Reversing React Native Apps?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Reversing React Native Apps?

As many as you like. As long as a model supports skills, you can use Reversing React Native Apps with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Reversing React Native Apps AI skill free?

Yes. It is published on GitHub by trilwu under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇