The OWASP Foundation
[Link]
Security Testing Guidelines
for mobile Apps
Florian Stahl
Johannes Strher
AppSec Research EU 2013
Who we are
Florian Stahl
Johannes Strher
Lead Consultant for Information
Security, CISSP, CIPP/IT
Consultant for Information
Security
Security & Privacy advocate
Expert for Mobile App Testing
Works in Munich for msg
systems ag, Germanys 5th
largest IT consulting and system
development company
Developed the Mobile Security
Testing Guide in his Masters
Thesis
[Link]@[Link]
[Link]@[Link]
AppSec Research EU 2013
Agenda
1. Motivation for Mobile Security Testing Guidelines
Current mobile threat landscape and current situation
Challenges
2. Mobile Security Testing Guide (MSTG)
Overview
Intelligence Gathering, Threat Modeling & Vulnerability Analysis in specific
Tools and examples
3. Summary
AppSec Research EU 2013
Mobile App Threat Landscape
Location-independent (mobile)
Always online and traceable
Consumerization devices are built for personal use
Focus on functionality and design rather than security
Raise of sensitive use cases for mobile apps
163% increase of mobile malware in 2012 *
Hidden business cases for free apps
* Source: NQ Mobile Security Report
AppSec Research EU 2013
Situation Mobile Security Testing
Mobile apps have some specific characteristics
regarding penetration testing
Custom guidelines have not been available
msg systems decided to develop guidelines (MSTG)
with Munich University of Applied Sciences
Similar guidelines published by OWASP:
OWASP Mobile Security Testing
AppSec Research EU 2013
Challenges
Identify differences to common penetration tests
Flexible Preconditions
App Security also depends on device security (jailbreak, different
platforms, versions, interfaces, MDM, etc.)
Different attackers (internal, external, network or device access,
blackbox / whitebox, etc.)
Keep it flexible AND give specific hints to the
penetration tester
Result: General process (mandatory) and supporting
tools and practices (optional)
AppSec Research EU 2013
Mobile Security Testing Guide
Overview
Start
Preparation
Intelligence Gathering
Threat Modeling
Vulnerability Analysis
Vulnerability
Assessment
Yes
Develop
Countermeasures
Show Countermeasures?
Explanation:
No
End
Mandatory
Optional
AppSec Research EU 2013
Annotation for
app specific sub-processes
The specific sub-processes were elaborated in detail
for Android and iOS
An iOS native CRM app is used for illustration
because
The CRM app supports many testable functions (authentication, )
It is open source more possibilities to demonstrate static
methods
It is a native app provides more attack surface for the tester
We can install the relating CRM service on an own server no
need for taking care of impacts during the tests
The CRM App was tested on an iPhone 4 with iOS 6
AppSec Research EU 2013
Intelligence Gathering
Try to catch as much as possible information about
the app
Intelligence
Gathering
Consists of 2 analysis
Environmental
Analysis
Architectural
Analysis
Differences to conventional process
Focus mainly on the architectural/technical part
Not considering mobile specific requirements
AppSec Research EU 2013
Intelligence Gathering
Environmental Analysis
Focus on the company behind the app and their business case and
the relating stakeholders
Analyze internal processes and structures
Architectural Analysis
App (network interfaces, used data, communication with other
ressources, session management, jailbreak/rooting detection, )
Runtime environment (MDM, jailbreak/rooting, os version)
Backend services (application server, databases, firewall, )
AppSec Research EU 2013
10
Intelligence Gathering - Example
Examples for collected information from the
Architectural Analysis for the CRM app
App
User session remains until the user logs off manually
No financial transactions are included
Runs on a jailbroken device no jailbreak detection
Provides operations on server side CRM data for creating, reading, updating,
deleting contacts, cases, calls,
Runtime environment analysis is not relevant, because the app is
running on a device from the tester
Backend services
Details about the version of the running CRM service
AppSec Research EU 2013
11
Threat Modeling
Identifying threats for the app - specific or
prepared threats (e. g. OWASP Top 10)
Should be done already in the development
Risk rating e. g. with OWASP Risk Rating
Developing countermeasures e. g. with
best practices or developers guides
Differences to conventional process
Most software testing processes do not include Threat
Modeling
Threat Modeling makes the complete process more
traceable and efficient for all participants
AppSec Research EU 2013
Start
Dividing and clustering
app
Identifying threats for
each component
Comparing all identified
threats
Rating all risks
Developing
countermeasures for
identified risks
Defining test cases
End
12
Threat Modeling - Example
Threat Modeling process example for the CRM App
Information from the Intelligence Gathering
App provides operations on CRM data on server side
Specific threat
Unauthorized reading of CRM data on the network traffic while
communicating with the CRM backend
Relating countermeasure
Implementing a secure transport layer protection (e. g. SSL, TLS)
Relating test case
Try to catch and read the network traffic between the CRM App and the
backend
AppSec Research EU 2013
13
Vulnerability Analysis
Identifying vulnerabilities in the app with the
previous created test cases
Executing test cases with techniques from 3 different
categories
Vulnerability
Analysis
Static
methods
Dynamic
methods
Forensic
methods
Differences to conventional process
Most software testing processes not include so many categories of
testing methods
AppSec Research EU 2013
14
Vulnerability Analysis
Static methods
Reverse Engineering
Automatic and manual source code analysis
Excursion: Tools for static methods
Reverse Engineering
Android: dex2jar, JD-GUI
iOS: otool, class-dump-z
Automatic and manual source code
analysis
Android: Androwarn, Andrubis,
ApkAnalyser
iOS: Flawfinder, Clang Static Analyzer
AppSec Research EU 2013
15
Vulnerability Analysis
Dynamic methods
Passive network monitoring and analyzing
Network traffic analysis at different places in the network (at the device,
gateway or in an own VPN)
Active network capturing and manipulating (Wifi and cellular)
Problems
Native apps do not use always device proxy settings
SSL encrypted connections
Solutions
Special apps that force the usage of device proxy settings or which break
SSL encrypted connections (mostly for jailbroken or rooted devices)
AppSec Research EU 2013
16
Vulnerability Analysis
Dynamic methods
Runtime analysis
Possible by analyzing the communicating process for internal components
(Android: Intents; iOS: objc_msgSend calls)
Runtime manipulation
Call or manipulate specific functions
Read and write variable values
File activity analysis
Analysis file system changes during the runtime
AppSec Research EU 2013
17
Vulnerability Analysis
Dynamic methods - CRM app example
Network traffic analysis reveals usage of HTTP and sending
non-encrypted sensitive user data (session id, username and
password)
Tools: Wireshark, BurpSuite,
User authentication can be bypassed by runtime manipulation
iOS tools: GNU debugger, Snoop-it, Cycript,
Android tools: Mercury, Intent Sniffer, Intent Fuzzer,
File activity analysis shows that user credentials (username and
password) are stored in and used from the iOS keychain
iOS tools: [Link], Snoop-it
Android tools: androidAuditTools
AppSec Research EU 2013
18
Vulnerability Analysis
Forensic methods
Timeline analysis
Analyze timestamps created from the file system
Analysis of different file types
SQLite databases
Screenshots (iOS)
Log files
Keyboard cache (iOS)
Cookies
SharedPreferences (Android)
Keychain (iOS)
AppSec Research EU 2013
19
Vulnerability Analysis
Forensic methods - CRM app example
Timeline analysis shows that the app updates several files during its
runtime (*.plist file, database)
Tools: mac-robber, mactime
Analyzing identified files and standard file types reveal that the user
credentials are stored in plain text in the iOS keychain
Tools: Keychain dumper, keychain viewer,
AppSec Research EU 2013
20
Summary
Mobile Security Testing Guide
considers mobile characteristics, but is independent
from technologies
helps to improve transparency and repeatability for
mobile penetration testing
is a holistic approach with sufficient flexibility
and ultimately helps to improve mobile app security
AppSec Research EU 2013
21
The OWASP Foundation
[Link]
Thank you for your attention!
infosec@[Link]
Full thesis (in German) available on request
AppSec Research EU 2013