SKILL.md
8.03 KB
title: Frida Android Setup and Hooking description: Install and configure frida on Android/Termux for dynamic instrumentation name: frida-android-setup version: 1.0 triggers:
- frida
- android hook
- dynamic instrumentation
- reverse engineering
Frida Android Setup and Hooking
Overview
Install frida-server and frida Python client on Android/Termux for dynamic instrumentation of Android apps.
Prerequisites
- Android device with root access (Magisk)
- Termux installed
- Internet connection for downloading binaries
Installation Steps
1. Install frida-server
# Download frida-server for Android arm64
VERSION=17.15.3
curl -L -o frida-server-${VERSION}-android-arm64.xz \
"https://github.com/frida/frida/releases/download/${VERSION}/frida-server-${VERSION}-android-arm64.xz"
# Extract
unxz frida-server-${VERSION}-android-arm64.xz
# Install to /data/local/tmp (requires root)
su -c "cp frida-server-${VERSION}-android-arm64 /data/local/tmp/frida-server"
su -c "chmod 755 /data/local/tmp/frida-server"
# Start frida-server
su -c "/data/local/tmp/frida-server -D"
# Verify
su -c "/data/local/tmp/frida-server --version"
2. Install frida Python Client
2.1 Download frida-core-devkit
VERSION=17.15.3
curl -L -o frida-core-devkit-${VERSION}-android-arm64.tar.xz \
"https://github.com/frida/frida/releases/download/${VERSION}/frida-core-devkit-${VERSION}-android-arm64.tar.xz"
# Extract to home directory
tar -xf frida-core-devkit-${VERSION}-android-arm64.tar.xz
# Produces: libfrida-core.a, frida-core.h, frida-core-example.c
2.2 Install Build Dependencies
pkg install clang make pkg-config glib libffi
pip install meson
2.3 Compile ninja (if not available)
# Download ninja source
curl -L -o ninja-src.tar.gz \
"https://github.com/ninja-build/ninja/archive/refs/tags/v1.12.1.tar.gz"
tar -xzf ninja-src.tar.gz
cd ninja-1.12.1
# Fix Android Bionic compatibility in src/util.cc
# The GetLoadAverage() function uses sysinfo/SI_LOAD_SHIFT which are not available on Android Bionic
# Replace the function with a simple return -0.0f
python3 configure.py --bootstrap
cp ninja $PREFIX/bin/
2.4 Download frida Python Source
pip download frida --no-deps -d frida_wheel
cd frida_wheel
tar -xzf frida-*.tar.gz
2.5 Compile frida Python Extension
mkdir frida_build && cd frida_build
# Create meson.build
cat > meson.build << 'EOF'
project('frida-python-ext', 'c', version: '17.15.3')
python = import('python').find_installation()
python_dep = python.dependency()
cc = meson.get_compiler('c')
frida_core_dep = cc.find_library('frida-core', dirs: '/data/data/com.termux/files/home')
frida_core_inc = include_directories('/data/data/com.termux/files/home')
glib_dep = dependency('glib-2.0', required: false)
gio_dep = dependency('gio-2.0', required: false)
os_deps = []
if glib_dep.found()
os_deps += glib_dep
endif
if gio_dep.found()
os_deps += gio_dep
endif
extension = python.extension_module('_frida',
'/path/to/frida-17.15.3/frida/_frida/extension.c',
limited_api: '3.7',
include_directories: frida_core_inc,
dependencies: [python_dep, frida_core_dep, os_deps],
install: true,
subdir: 'frida',
)
EOF
# Build
meson setup build
meson compile -C build
# The output will be at build/_frida.abi3.so
2.6 Install frida Python Package
cd frida-17.15.3
FRIDA_EXTENSION=/path/to/frida_build/build/_frida.abi3.so \
pip install . --no-build-isolation
2.7 Verify Installation
python3 -c "import frida; print(frida.__version__)"
Usage
Connect to frida-server
import frida
# Connect to local frida-server
device = frida.get_device_manager().add_remote_device('127.0.0.1:27042')
# List processes
processes = device.enumerate_processes()
for p in processes:
print(f'{p.pid}: {p.name}')
Attach to a Process
# Attach by process name
session = device.attach('com.example.app')
# Attach by PID
session = device.attach(1234)
Load a Script
script = session.create_script('''
Java.perform(function() {
console.log('Java VM available: ' + Java.available);
// Hook a Java method
var TargetClass = Java.use('com.example.app.TargetClass');
TargetClass.targetMethod.implementation = function(arg1, arg2) {
console.log('Hook called with: ' + arg1 + ', ' + arg2);
return this.targetMethod(arg1, arg2);
};
});
''')
script.on('message', lambda msg, data: print(msg))
script.load()
Pitfalls
Python 3.14 无预编译 wheel(Termux 2026-07 验证)
- Termux 默认 Python 升级到 3.14 后,
pip install frida和pip install frida-tools均失败 - PyPI 上 frida 没有 cp314 预编译 wheel,
--only-binary :all:也找不到匹配版本 - 源码编译需要 ninja + meson + glib + frida-core-devkit,在 Android 16 Bionic 上极易失败
- 替代方案(按推荐顺序):
- 使用 frida-server 的 CLI 模式 +
frida-inject二进制(不需要 Python 客户端) - 在另一台机器上运行 frida Python 客户端,通过网络连接 frida-server(
frida -H 127.0.0.1:27042) - 放弃 Frida,改用 UI 自动化方案(适用于只需触发 App 功能、不需要 hook 内部逻辑的场景)
- 安装 Termux 旧版 Python 3.12 并行使用(如有)
- 使用 frida-server 的 CLI 模式 +
Version Mismatch
- frida-server and frida Python client must be the same version
- Mismatched versions cause Java VM detection failures
Java VM Not Available
- On some Android versions (e.g., Android 16), Java VM may not be available after attach
- This is a known limitation; use native hooking instead
Android 16 / Bionic Compatibility
- Android 16 uses Bionic libc which lacks
sysinfostructure andSI_LOAD_SHIFT - ninja build fails with
sysinfoandSI_LOAD_SHIFTerrors - Fix: Patch
src/util.ccin ninja source to replaceGetLoadAverage()with simplereturn -0.0f - See
references/frida-termux-compilation-recipe.mdfor full patch
frida-server Version Mismatch Symptoms
- Java VM unavailable after attach
- Process list shows wrong process names (e.g., "mmd" instead of "com.tencent.mm")
- Fix: Ensure frida-server and frida Python client versions match exactly
- Download matching versions from GitHub releases
Process Name Resolution
- frida matches processes by human-readable label, not package name
- WeChat main process may show as "WeChat" or "mmd" instead of "com.tencent.mm"
- Multi-process apps may resolve to wrong sub-process
/data/local/tmp Permission
- frida needs write access to /data/local/tmp for helper files
- Run frida-server as root or ensure /data/local/tmp is writable
Frida 17.x Spawn Bug on Android 16
frida -f <package>(spawn mode) fails on Android 16:Failed to spawn: unable to locate android.os.Process.setArgV0() slot- This is a known issue with Frida 17.x on newer Android versions
- Workaround: Use attach mode (
frida -p <PID>) instead of spawn mode - Start app normally (
am start), wait, then attach by PID
Anti-Frida Detection (Xiaohongshu, banking apps, etc.)
- Apps like Xiaohongshu actively detect Frida. Symptoms:
Failed to attach: agent connection closed unexpectedly - The app detects the frida agent in-process and kills the connection
- Attempted bypasses (limited success):
- Rename frida-server binary:
cp frida-server fsrv - Change port:
fsrv -l 0.0.0.0:31337 - Attach with:
frida -H 127.0.0.1:31337 -p <PID>
- Rename frida-server binary:
- These may still fail if the app checks for Frida's in-process artifacts
- For apps with strong anti-Frida, prefer static analysis (jadx, apktool, MMKV extraction)
References
- frida releases
- frida docs
- apktool-android-apk-decompilation
- wechat-send-message-mechanism
- wechat-internal-send-api — sk.p0() method analysis and Frida hook target
- frida-termux-compilation-recipe