dragonjar-android-pentesting-skill

---

name: dragonjar-android-pentesting-skill description: Comprehensive Android APK security analysis with static/dynamic testing, RASP detection, Frida instrumentation, and MASVS compliance scoring triggers:

• audit this Android APK for security vulnerabilities
• analyze this APK with MASVS scoring and runtime defense analysis
• bypass SSL pinning and root detection in this Android app
• detect RASP protections and generate Frida bypass scripts
• decompile this APK and find hardcoded secrets
• run static and dynamic analysis on this Android application
• trace data flows and validate security controls in this APK
• generate a pentesting report for this Android app

---

DragonJAR Android Pentesting Skill

Skill by ara.so — Security Skills collection.

This skill provides comprehensive Android APK security analysis capabilities for AI agents, combining static analysis, dynamic instrumentation with Frida, RASP detection, authorized bypass validation, source-to-sink tracing, MASVS scoring, and professional reporting in a unified workflow.

What This Skill Does

Transforms an AI agent into an expert Android security auditor capable of:

• APK Analysis: Decode APKs with APKTool, decompile with JADX, detect frameworks with APKiD
• Static Security Analysis: 50+ manifest checks, 70+ Semgrep MASTG rules, secret detection, obfuscation analysis
• Dynamic Instrumentation: 37 Frida scripts for SSL pinning bypass, root detection bypass, crypto interception
• Runtime Defense Analysis (RDA): Detect 18 protection categories (RootBeer, SafetyNet, Frida detection, RASP, etc.)
• RASP Bypass: Authorized bypass runner with reusable profiles, DRY workflow
• Data Flow Tracing: Source-to-sink methodology with confidence levels
• MASVS Compliance: Automated scoring against OWASP MASVS controls with CVSS 4.0
• APK Modification: Smali patching, repackaging, signing, validation

Installation

Prerequisites

Install required tools (Linux/macOS):

# APKTool (3.0.1+)
brew install apktool

# JADX (1.5.5+)
brew install jadx

# Android SDK Platform Tools
brew install --cask android-platform-tools

# Frida (17.9+)
pip3 install frida-tools

# Objection (1.12.4+)
pip3 install objection

# APKiD (3.0.0+)
pip3 install apkid

# Semgrep (optional, for SAST)
brew install semgrep

Skill Installation

# Clone into your agent's skills directory
cd ~/.agents/skills/
git clone https://github.com/DragonJAR/Android-Pentesting-Skill dragonjar-android-pentesting

# Or clone to any location and add to agent skill path
git clone https://github.com/DragonJAR/Android-Pentesting-Skill.git

Verification

# Verify all dependencies are installed
cd dragonjar-android-pentesting
python3 scripts/06-setup/preflight-check.py

# Expected output:
# ✅ APKTool 3.0.1 found
# ✅ JADX 1.5.5 found
# ✅ Frida 17.9.4 found
# ✅ All dependencies satisfied

Core Workflows

1. Basic APK Security Audit

# Full static analysis with Semgrep enrichment
bash scripts/auto-audit-static.sh /path/to/app.apk --semgrep

# Output includes:
# - Manifest security issues (exported components, permissions)
# - Hardcoded secrets (API keys, passwords, tokens)
# - Insecure crypto usage
# - WebView vulnerabilities
# - Intent injection risks
# - Findings report in findings-merged.json

Example findings output:

{
  "findings": [
    {
      "id": "HARD-001",
      "severity": "CRITICAL",
      "title": "Hardcoded API Key in Source Code",
      "owasp_mobile": "M1",
      "masvs_control": "MASVS-STORAGE-1",
      "cvss": "9.1 (CRITICAL)",
      "location": "com/example/app/Config.java:42",
      "evidence": "private static final String API_KEY = \"sk_live_...\";",
      "remediation": "Store API keys in BuildConfig or secure server-side configuration"
    }
  ]
}

2. Runtime Defense Analysis (RASP Detection)

# Detect runtime protections with passive + active modes
bash scripts/02-rasp/runtime-defense-analyzer.sh \
  /path/to/app.apk \
  com.example.app \
  --active-mode \
  --authorized-lab \
  --output findings-rda.json

# Catalog includes 18 protection categories:
# - Root detection (RootBeer, custom native)
# - Emulator detection
# - Debug detection
# - Frida/instrumentation detection
# - Screenshot/screenrecord blocking
# - SafetyNet / Play Integrity
# - Commercial RASP (Talsec, AppSealing, DexGuard, etc.)

RDA output structure:

{
  "detectors": {
    "rootbeer": {
      "status": "DETECTED",
      "confidence": "high",
      "evidence": "RootBeer.isRooted() returns true",
      "bypass_profile": "rootbeer_standard"
    },
    "ssl_pinning": {
      "status": "DETECTED",
      "implementation": "OkHttp3 CertificatePinner",
      "bypass_profile": "ssl_okhttp3"
    }
  }
}

3. RASP Bypass Workflow (DRY Pattern)

# 1) List available bypass profiles
bash scripts/02-rasp/rasp-bypass-runner.sh --list-profiles

# Output:
# Available profiles:
# - rootbeer_standard (RootBeer library bypass)
# - ssl_okhttp3 (OkHttp3 SSL pinning)
# - ssl_trustmanager (TrustManager bypass)
# - frida_detection (Anti-Frida bypass)
# - emulator_detection (Emulator checks bypass)

# 2) Generate bypass command from RDA findings (print only)
bash scripts/02-rasp/rasp-bypass-runner.sh \
  --package com.example.app \
  --from-rda findings-rda.json \
  --print-command

# Output:
# frida -U -f com.example.app \
#   -l assets/frida-scripts/android-root-bypass-advanced.js \
#   -l assets/frida-scripts/ssl-pinning-bypass.js \
#   --no-pause

# 3) Execute bypass in authorized lab environment
bash scripts/02-rasp/rasp-bypass-runner.sh \
  --package com.example.app \
  --from-rda findings-rda.json \
  --run \
  --authorized-lab

# Launches Frida with combined bypass scripts

Important: Client-side bypasses do NOT forge server-side attestation. For Play Integrity, SafetyNet, Approov, or similar backend-enforced controls, use an authorized test tenant, backend allowlist, or approved lab configuration.

4. SSL Pinning Bypass

# Universal SSL pinning bypass (30+ implementations)
python3 scripts/07-tools/frida-exploit-helper.py \
  -p com.example.app \
  --script ssl-pinning-bypass

# Supports:
# - OkHttp3 CertificatePinner
# - TrustManager custom implementations
# - WebView SSL error handlers
# - React Native ssl-pinning libraries
# - Flutter BoringSSL (native hooks)
# - Cordova SSL plugins

5. Framework-Specific Analysis

React Native

# Detect React Native and extract JavaScript bundle
bash scripts/auto-audit-static.sh /path/to/app.apk

# Automatically:
# - Detects libreactnativejni.so
# - Extracts assets/index.android.bundle
# - Scans bundle for hardcoded secrets
# - Analyzes Metro bundler output

Hook React Native bridge:

// assets/frida-scripts/react-native-bridge-hook.js
Java.perform(function() {
    var CatalystInstanceImpl = Java.use('com.facebook.react.bridge.CatalystInstanceImpl');
    
    CatalystInstanceImpl.jniCallJSFunction.implementation = function(module, method, args) {
        console.log('[RN Bridge] ' + module + '.' + method);
        console.log('[RN Bridge] Args: ' + JSON.stringify(args));
        return this.jniCallJSFunction(module, method, args);
    };
});

Flutter

# Flutter uses Dart AOT compilation and native BoringSSL
# Standard Java SSL hooks won't work

# Use Blutter for Dart code extraction
python3 tools/blutter/blutter.py lib/arm64-v8a/libapp.so output/

# Hook native SSL functions
python3 scripts/07-tools/frida-exploit-helper.py \
  -p com.example.flutter \
  --script flutter-ssl-bypass

# Hooks:
# - SSL_CTX_set_custom_verify (certificate validation)
# - SSL_read / SSL_write (traffic interception)

6. Data Flow Tracing

# Phase 3: Source-to-sink analysis
# Automatically runs during static audit

# Example traced flows:
# - User input → SQL query (SQL injection risk)
# - Intent extras → WebView.loadUrl (open redirect)
# - SharedPreferences → network (data leakage)
# - File paths → external storage (path traversal)

Flow confidence levels:

• CONFIRMED: Direct observable flow without conditions
• LIKELY: Flow with minimal conditional branches
• POSSIBLE: Flow through complex logic, needs dynamic validation

7. MASVS Compliance Scoring

# Calculate OWASP MASVS compliance score
python3 scripts/05-reporting/masvs-scorer.py findings-merged.json

# Output:
# MASVS Score: 72.6/100 (Grade: C)
# Controls Passed: 16/23
# Controls Failed: 7
#
# Critical Failures:
# - MASVS-STORAGE-1: Hardcoded secrets in code
# - MASVS-CRYPTO-1: Weak cryptographic algorithms
#
# Recommendations:
# - Implement secure credential storage (Android Keystore)
# - Upgrade to AES-256-GCM for encryption

8. APK Modification Workflow

# Decode APK
apktool d -f -o decoded/ /path/to/app.apk

# Modify smali code (example: disable root check)
# Edit decoded/smali/com/app/RootDetector.smali
# Change: const/4 v0, 0x1  (return true)
# To:     const/4 v0, 0x0  (return false)

# Rebuild APK
apktool b decoded/ -o app-modified.apk

# Align and sign
zipalign -v -p 4 app-modified.apk app-aligned.apk
apksigner sign --ks ~/.android/debug.keystore \
  --ks-pass pass:android \
  --out app-signed.apk \
  app-aligned.apk

# Verify signature
apksigner verify --verbose app-signed.apk

Frida Script Library

Common Frida Operations

# List all available Frida scripts
ls assets/frida-scripts/

# Key scripts:
# - ssl-pinning-bypass.js (universal SSL bypass)
# - android-root-bypass-advanced.js (30+ root detection bypasses)
# - crypto-intercept.js (monitor Cipher, MessageDigest, etc.)
# - biometric-bypass.js (BiometricPrompt, FingerprintManager)
# - keystore-inspector.js (dump Keystore entries)
# - webview-inspector.js (WebView debugging, JS injection)
# - intent-fuzzer.js (Intent injection testing)

Frida Exploit Helper

# List bundled scripts
python3 scripts/07-tools/frida-exploit-helper.py --list-scripts

# Hook memory functions
python3 scripts/07-tools/frida-exploit-helper.py \
  -p com.example.app \
  --hook malloc,free,memcpy

# SSL pinning bypass
python3 scripts/07-tools/frida-exploit-helper.py \
  -p com.example.app \
  --script ssl-pinning-bypass

# Memory layout analysis
python3 scripts/07-tools/frida-exploit-helper.py \
  -p com.example.app \
  --layout

# Runtime defense detection
python3 scripts/07-tools/frida-exploit-helper.py \
  -p com.example.app \
  --runtime-defense

Custom Frida Script Example

// Hook custom native function
Java.perform(function() {
    var targetClass = Java.use('com.example.app.SecurityCheck');
    
    targetClass.isDeviceSecure.implementation = function() {
        console.log('[+] isDeviceSecure() called');
        var result = this.isDeviceSecure();
        console.log('[+] Original result: ' + result);
        console.log('[+] Forcing return: true');
        return true;
    };
    
    console.log('[+] Hooked isDeviceSecure()');
});

Configuration

Environment Variables

# Android SDK path
export ANDROID_HOME="$HOME/Library/Android/sdk"
export PATH="$PATH:$ANDROID_HOME/platform-tools:$ANDROID_HOME/build-tools/36.0.0"

# Frida server configuration
export FRIDA_SERVER_PORT=27042

# Optional: Semgrep API token for managed scanning
export SEMGREP_APP_TOKEN="your_token_here"

# Optional: Custom APKTool config
export APKTOOL_CONFIG="$HOME/.apktool/config.yml"

Bypass Profiles Configuration

// scripts/02-rasp/bypass-profiles.json
{
  "profiles": {
    "rootbeer_standard": {
      "description": "RootBeer library bypass",
      "scripts": [
        "assets/frida-scripts/android-root-bypass-advanced.js"
      ],
      "hooks": ["RootBeer.isRooted", "RootBeer.isRootedWithoutBusyBoxCheck"]
    },
    "ssl_okhttp3": {
      "description": "OkHttp3 CertificatePinner bypass",
      "scripts": [
        "assets/frida-scripts/ssl-pinning-bypass.js"
      ],
      "hooks": ["CertificatePinner.check"]
    }
  }
}

Common Patterns

Pattern 1: Full Security Assessment

# 1) Static analysis with Semgrep
bash scripts/auto-audit-static.sh app.apk --semgrep

# 2) Runtime defense detection
bash scripts/02-rasp/runtime-defense-analyzer.sh \
  app.apk com.example.app \
  --active-mode --authorized-lab \
  --output findings-rda.json

# 3) MASVS scoring
python3 scripts/05-reporting/masvs-scorer.py findings-merged.json

# 4) Generate professional report
python3 scripts/05-reporting/report-generator.py \
  findings-merged.json \
  findings-rda.json \
  --output report.pdf

Pattern 2: Traffic Interception

# 1) Set up proxy (Burp/ZAP)
export HTTP_PROXY=127.0.0.1:8080
export HTTPS_PROXY=127.0.0.1:8080

# 2) Install CA certificate on device
adb push burp-ca.crt /sdcard/
# Install via Settings → Security → Install from SD card

# 3) Bypass SSL pinning
frida -U -f com.example.app \
  -l assets/frida-scripts/ssl-pinning-bypass.js \
  --no-pause

# 4) Monitor traffic in proxy

Pattern 3: Root Detection Bypass Stack

# Layer 1: Java-level root checks
frida -U -f com.example.app \
  -l assets/frida-scripts/android-root-bypass-advanced.js \
  --no-pause

# Layer 2: Native-level root checks (if detected)
# Edit native-root-bypass.js to target specific native libraries
frida -U -f com.example.app \
  -l assets/frida-scripts/android-root-bypass-advanced.js \
  -l assets/frida-scripts/native-hook.js \
  --no-pause

Pattern 4: Automated Secret Extraction

# Extract and scan all strings
bash scripts/01-decompile/extract-strings.sh app.apk > strings.txt

# Scan with patterns
grep -E '(sk_live_|ghp_|AIza[0-9A-Za-z-_]{35})' strings.txt

# Deep search in decompiled code
find decoded/smali -name "*.smali" -exec grep -H "const-string.*sk_live" {} \;

# Search in JavaScript bundles (React Native/Cordova)
find decoded/assets -name "*.bundle" -o -name "*.js" | \
  xargs grep -E '(API_KEY|SECRET|PASSWORD).*=.*["\'][^"\']{20,}'

Troubleshooting

Issue: APKTool Decode Fails

# Error: "brut.androlib.AndrolibException: Could not decode arsc file"

# Solution 1: Update APKTool to 3.0.1+
brew upgrade apktool

# Solution 2: Use --only-main-classes flag
apktool d --only-main-classes -f -o decoded/ app.apk

# Solution 3: Use legacy AAPT mode (not recommended)
apktool d --use-aapt1 -f -o decoded/ app.apk

Issue: Frida Server Connection Failed

# Error: "Failed to spawn: unable to find process with name 'com.example.app'"

# Solution 1: Verify Frida server is running
adb shell "su -c '/data/local/tmp/frida-server &'"

# Solution 2: Check Frida server version matches client
frida --version  # Client version
adb shell "/data/local/tmp/frida-server --version"  # Server version

# Solution 3: Use spawn mode instead of attach
frida -U -f com.example.app  # Spawn mode
# Instead of:
frida -U com.example.app      # Attach mode

Issue: SSL Pinning Bypass Not Working

# Flutter apps use native BoringSSL - Java hooks won't work
# Solution: Use Flutter-specific native hooks
frida -U -f com.example.flutter \
  -l assets/frida-scripts/flutter-ssl-bypass.js

# React Native apps may use custom native modules
# Solution: Hook both Java and native SSL functions
frida -U -f com.example.rn \
  -l assets/frida-scripts/ssl-pinning-bypass.js \
  -l assets/frida-scripts/native-hook.js

Issue: Runtime Defense Analysis Returns Empty Results

# Passive mode only checks static indicators
# Solution: Use --active-mode with authorized lab

bash scripts/02-rasp/runtime-defense-analyzer.sh \
  app.apk com.example.app \
  --active-mode \
  --authorized-lab \
  --output findings-rda.json

# Ensure device/emulator is accessible
adb devices  # Should show device

# Ensure app is debuggable or use root
adb shell "su -c 'pm list packages | grep example'"

Issue: MASVS Scorer Shows Low Score

# Low scores indicate security gaps - this is expected behavior
# Review failed controls:

python3 scripts/05-reporting/masvs-scorer.py findings-merged.json --verbose

# Common failure patterns:
# - MASVS-STORAGE-1: Hardcoded secrets → Use Android Keystore
# - MASVS-CRYPTO-1: Weak crypto → Upgrade to AES-256-GCM
# - MASVS-NETWORK-1: No SSL pinning → Implement certificate pinning
# - MASVS-RESILIENCE-1: No obfuscation → Apply ProGuard/R8

Agent Usage Examples

Example 1: Quick APK Audit

User: "Audit this APK: ~/Downloads/banking-app.apk"

Agent Actions:

cd dragonjar-android-pentesting
bash scripts/auto-audit-static.sh ~/Downloads/banking-app.apk --semgrep
python3 scripts/05-reporting/masvs-scorer.py findings-merged.json

Agent Response: "Found 8 security issues: 2 CRITICAL (hardcoded API keys), 3 HIGH (exported activities without permissions), 2 MEDIUM (weak crypto), 1 LOW (debug logs). MASVS score: 68.2/100 (Grade: D). Report saved to findings-merged.json."

Example 2: Bypass SSL Pinning

User: "This banking app has SSL pinning. I need to intercept the login request."

Agent Actions:

# Start Frida with SSL bypass
python3 scripts/07-tools/frida-exploit-helper.py \
  -p com.banking.app \
  --script ssl-pinning-bypass

Agent Response: "SSL pinning bypassed (OkHttp3 CertificatePinner detected). Configure your proxy (Burp/ZAP) to 127.0.0.1:8080. Restart the app and attempt login."

Example 3: Detect RASP Protections

User: "Check what runtime protections this app has."

Agent Actions:

bash scripts/02-rasp/runtime-defense-analyzer.sh \
  ~/Downloads/app.apk com.protected.app \
  --active-mode --authorized-lab \
  --output findings-rda.json
cat findings-rda.json

Agent Response: "Detected 4 protections: RootBeer (root detection), custom emulator checks, Frida detection (anti-instrumentation), DexGuard obfuscation. Bypass profiles available: rootbeer_standard, frida_detection, emulator_detection. Run rasp-bypass-runner.sh to generate bypass command."

Example 4: Find Hardcoded Secrets in React Native

User: "This is a React Native app. Find any hardcoded API keys in the JavaScript bundle."

Agent Actions:

bash scripts/auto-audit-static.sh ~/Downloads/rn-app.apk
grep -r "API_KEY\|api_key\|apiKey" decoded/assets/*.bundle

Agent Response: "Found 3 hardcoded secrets in assets/index.android.bundle: Firebase API key (line 4521), Stripe publishable key (line 8912), AWS access key ID (line 12045). Evidence saved to findings-merged.json with MASVS-STORAGE-1 control failure."

Advanced Topics

Native Code Analysis

# Extract native libraries
unzip app.apk "lib/*" -d native/

# Analyze with Ghidra/IDA (manual)
# Or use Frida for runtime analysis:

python3 scripts/07-tools/frida-exploit-helper.py \
  -p com.example.app \
  --hook JNI_OnLoad,RegisterNatives

# Hook specific native function by offset
frida -U -f com.example.app -l - << 'EOF'
var base = Module.findBaseAddress('libnative.so');
Interceptor.attach(base.add(0x1234), {
    onEnter: function(args) {
        console.log('[+] Native function called');
        console.log('Arg0: ' + args[0]);
    }
});
EOF

Attestation Bypass Limitations

**Important**: Client-side hooks cannot forge server-verified attestation:

- ❌ Play Integrity verdicts (Google server-signed)
- ❌ SafetyNet attestation responses
- ❌ Approov tokens (server-side verification)
- ❌ App Attest (Apple server validation)

**Authorized Testing Approaches**:
- ✅ Use test tenant with backend allowlist
- ✅ Configure mock verifier in staging environment
- ✅ Request official pentest exception from vendor
- ✅ Use approved lab environment with vendor cooperation

Custom Semgrep Rules

# Add custom rule to scripts/03-static-analysis/semgrep-rules/

rules:
  - id: custom-api-key-pattern
    pattern: |
      const-string $VAR, "cust_$KEY"
    message: Custom API key pattern detected
    severity: ERROR
    languages: [smali]
    metadata:
      owasp_mobile: M1
      masvs_control: MASVS-STORAGE-1

References

• OWASP MASVS: https://mas.owasp.org/MASVS/
• OWASP MASTG: https://mas.owasp.org/MASTG/
• Frida Documentation: https://frida.re/docs/
• APKTool Documentation: https://apktool.org/docs/
• DragonJAR Community: https://www.dragonjar.org/

License

Apache 2.0 - See LICENSE file for details.