The OWASP Mobile Application Security Testing Guide (MASTG), currently v2.0.0 (released 30 June 2026), is, per OWASP's own MASVS documentation, the resource that "describes the technical processes for verifying the controls listed in the MASVS." It is maintained by the OWASP Mobile Application Security (MAS) project and, per its GitHub description, serves as "a comprehensive manual for mobile app security testing and reverse engineering" covering Android and iOS apps. Where MASVS is the checklist of what to verify, MASTG is the how: concrete test procedures, tooling, and worked demonstrations for each control.
Key takeaways
- MASTG's structure changed substantially with v2.0.0, released 30 June 2026.
- For years, MASVS and MASTG stood as a two-piece system: control, then test.
- You trace it down three layers.
- Every v2 test page has the same four parts: Overview, Steps, Observation and Evaluation.
A modular guide, not a single document
MASTG's structure changed substantially with v2.0.0, released 30 June 2026. Per the project's GitHub release notes, this release completed a multi-year refactor that broke the guide out of large per-topic markdown pages into individually referenceable, machine-readable components, each with its own stable identifier and cross-links to the MASVS control and MASWE weakness it addresses:
| Component type | Count at v2.0.0 |
|---|---|
Tests (MASTG-TEST-####) |
193 |
| Techniques | 167 |
| Knowledge-base articles | 140 |
| Best practices | 72 |
| Demos | 152 |
| Tools | 135 |
These figures are a snapshot of the v2.0.0 release specifically — the guide is under continuous development, so current totals should be read on the live site rather than assumed to stay fixed at these numbers.
MASWE: the missing middle layer
For years, MASVS and MASTG stood as a two-piece system: control, then test. MASWE (Mobile Application Security Weakness Enumeration) fills the gap between them with a catalog of specific, named weaknesses — for example MASWE-0001, "Sensitive Data Stored Unencrypted in Private Storage" — each mapped back to the MASVS control it violates and forward to the MASTG test(s) that can detect it. OWASP announced MASWE v1.0.0 in a MAS-project blog post dated 17 August 2026, "MASWE v1.0.0 is Here." Read together, the three pieces now form a chain: MASVS states the control, MASWE names the concrete failure mode, and MASTG supplies the test that proves whether that failure is present in a given app.
How does one requirement become a MASTG test verdict?
You trace it down three layers. The v2.0.0 release notes describe the path as MASVS control, then MASWE weakness, then MASTG test, with demos as runnable examples. Here is that chain for data in transit, as listed on mas.owasp.org on 2026-11-09:
| Layer | Item | What it says |
|---|---|---|
| MASVS control | MASVS-NETWORK-1 | "The app secures all network traffic according to the current best practices." Related weaknesses: MASWE-0026, -0027, -0047, -0062 |
| MASWE weakness | MASWE-0026: Network Traffic Not Encrypted | Maps to CWE-319 (Cleartext Transmission of Sensitive Information) |
| MASTG tests (examples) | MASTG-TEST-0235 (Android configuration), MASTG-TEST-0236 (cleartext observed on the network), MASTG-TEST-0322 (iOS App Transport Security configuration) | Each test is one platform-specific check with its own pass/fail rule |
A tester can therefore start from the control and land on platform-specific checks with defined fail conditions. A finding can then cite both the MASWE ID and the CWE. That is useful when the same product's firmware findings are already reported against CWE.
What does a single MASTG test look like?
Every v2 test page has the same four parts: Overview, Steps, Observation and Evaluation. MASTG-TEST-0235: Android App Configurations Allowing Cleartext Traffic is a good example. Its overview notes that cleartext HTTP is blocked by default since Android 9 (API level 28), but an app can still enable it in two places. The steps are:
- Reverse engineer the app package.
- Obtain the
AndroidManifest.xml. - Read
android:usesCleartextTrafficand check whetherandroid:networkSecurityConfigis present. - Read
cleartextTrafficPermittedin the<base-config>and<domain-config>elements of the Network Security Configuration.
Each step links to a reusable MASTG technique, so the same "obtain the manifest" procedure is written once and shared by many tests.
When the test fails. Per the test page, cleartext is permitted, and the test fails, if the manifest sets
usesCleartextTrafficto true with no Network Security Configuration, or if the configuration setscleartextTrafficPermittedto true in<base-config>or in any<domain-config>. If a Network Security Configuration exists, even an empty one, the manifest flag alone does not fail the test.
That explicit evaluation rule is what makes a MASTG-based result repeatable: two testers looking at the same APK should reach the same verdict.
Which MAS profile should your app be tested against?
Pick the profile from the attacker you need to resist. The MAS Testing Profiles page defines four defaults, and each test on the tests index is tagged with the profiles it belongs to:
| Profile | Attacker model (OWASP wording, condensed) | OWASP's example apps |
|---|---|---|
| MAS-L1 (Essential Security) | Other apps installed on the device are adversaries | News, calendar |
| MAS-L2 (Advanced Security) | The operating system cannot be trusted; attackers may have physical access to the device | Messenger, health, sport; data such as location, payment, health, access tokens, API keys |
| MAS-R (Resilient Security) | The device user is the attacker, including reverse engineers and cheaters | Combined with L1 or L2 for apps with business assets, such as banking or games |
| MAS-P (Baseline Privacy) | Not attacker-centric; protects users' personal data | Combined with the others |
OWASP also publishes a specialized MAS-EUDIW profile for the EU Digital Identity Wallet. Its Using MAS Profiles page says the goal "is not to comply with every profile or every test." Choose from the app's threat model, features and data sensitivity, and prefer a custom profile built from a threat model when you can. The profile choice changes the test list. On 2026-11-09, sandbox-storage tests such as MASTG-TEST-0207 and MASTG-TEST-0287 were tagged L2 only, so an L1 assessment would not run them.
Are the old v1 MASTG tests still valid?
No. According to the v2.0.0 release notes, v1 tests remain on the website for a limited time behind a "Show Deprecated" toggle. They are no longer maintained, and the v2 tests are now the canonical reference. A deprecated v1 page shows a banner that links to its v2 replacements. One v1 port can split into several atomic tests; the release notes cite the biometric-authentication port, which produced 5 tests and 5 demos. On 2026-11-09 the tests index tagged 92 entries as deprecated. A test report that cites v1 IDs such as MASTG-TEST-0001 points to retired procedures.
Scope: native apps first
Per the MASTG's own overview, the guide focuses primarily on native Android and iOS apps rather than attempting exhaustive coverage of every cross-platform framework, on the basis that the underlying techniques — storage inspection, traffic interception, binary analysis, runtime instrumentation — carry over to hybrid and web-based mobile apps even when the framework-specific tooling differs. Teams testing a Flutter, React Native, or WebView-heavy app should expect to adapt MASTG's native test procedures rather than find a framework-specific walkthrough for every case.
License and OWASP's stance on certification
The MASTG is free to read at mas.owasp.org and licensed, per the OWASP/mastg GitHub repository, under Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) — the same license OWASP applies to MASVS. As with MASVS, OWASP does not certify vendors, testers, or software against the MASTG; any third party's claim of running "MASTG-based" or "MASTG-certified" testing is that party's own characterization of their process, not an OWASP-issued credential.
Where does this fit for device makers?
A companion mobile app for setup, pairing, or firmware updates is now standard on most connected devices, and it is frequently assessed far less rigorously than the firmware or hardware it controls. Teams that already use MASVS as the control baseline for that companion app can use MASTG as the concrete test plan: pick the MASVS controls relevant to the app's MAS Profile, then follow the matching MASTG tests rather than inventing ad hoc test steps. Firmware teams tracking CWE Top 25 weaknesses in embedded code gain a parallel, test-procedure-level resource for the mobile side of the same product.
SoftWSP's device and firmware assessments can extend to a product's companion mobile app, running it against MASTG-style test procedures rather than leaving the app as an unassessed gap next to a hardened firmware image: see device & firmware security services.
Worked example: what should a companion-app test plan cover?
Take a hypothetical connected door lock whose Android and iOS app handles pairing, cloud login and remote unlock. The app holds access tokens and device credentials. By OWASP's own profile examples, that points to MAS-L2 plus MAS-P. The plan below maps each app feature to IDs copied from the linked mas.owasp.org pages on 2026-11-09:
| App feature | MASVS control | MASWE weakness | MASTG tests to run |
|---|---|---|---|
| Cloud API calls | MASVS-NETWORK-1 | MASWE-0026 Network Traffic Not Encrypted | 0235 (Android), 0322 (iOS), 0236 (network capture) |
| Server identity checks | MASVS-NETWORK-1 | MASWE-0027 Insecure Certificate Validation | 0282, 0283 (Android), 0396 (iOS) |
| Stored session tokens and device keys | MASVS-STORAGE-1/-2, MASVS-CRYPTO-2 | MASWE-0001 Sensitive Data Stored Unencrypted in Private Storage | 0207, 0287 (Android), 0299 (iOS) |
| Debug and crash logging | MASVS-STORAGE-2 | MASWE-0005 Insertion of Sensitive Data into Logs | 0203, 0231 (Android), 0296 (iOS) |
| API keys inside the app package | MASVS-STORAGE-1 | MASWE-0004 Sensitive Data Hardcoded in the App Package | None listed on 2026-11-09; cover by manual review and record the gap |
Two parts of a device product sit outside this table. First, the Bluetooth pairing channel between phone and lock: test it against the device side as well, using our Bluetooth LE security guide. Second, the cloud backend. The MASVS assumes that backend servers, IoT and companion devices follow their own industry standards, such as ETSI EN 303 645 for consumer IoT.
What are the common pitfalls?
- Calling it certification. OWASP certifies no one against MASTG; describe the scope as "tested against MASTG v2 tests for profiles L2 + P" and list the test IDs.
- Citing retired v1 tests. v1 tests are deprecated and no longer maintained (release notes).
- Running every test. OWASP's profile guidance says to skip tests for features the app lacks, and to record why.
- Assuming MASWE coverage equals MASTG coverage. Some weaknesses, such as MASWE-0004 on 2026-11-09, list no MASTG test, so they need manual review.
- Testing one platform. Weaknesses such as MASWE-0026 have separate Android and iOS tests, each with its own evaluation rule.
- Skipping the threat model. The MASVS assumes secure architecture, design and threat modeling have been done first. Our threat modeling guide covers that step.
How do you run a MASTG-based assessment, step by step?
- Threat-model the app and its device ecosystem, then pick a MAS profile or a custom one (MAS Profiles).
- List the MASVS controls in scope and the MASWE weaknesses linked from each control page.
- From each MASWE page, select the current v2 tests for both platforms, filtered by your profile tags.
- Run each test's Steps, keep the Observation output, and apply its Evaluation rule as written.
- Record weaknesses that have no MASTG test, and how you covered them.
- Report each finding with its MASTG-TEST, MASWE and CWE IDs, the MASTG version used, and a CVSS score if your process needs one (see the CVSS calculator).
- Re-test after fixes against the same test IDs, so the before and after results can be compared.
The firmware side of the same product follows the OWASP Firmware Security Testing Methodology, which plays the same role for the embedded image.
This guide summarises a public OWASP publication; it does not constitute legal or compliance advice.
Put this guide to work
Frequently asked questions
What is the OWASP MASTG?
The Mobile Application Security Testing Guide (MASTG) is OWASP's companion resource to the Mobile Application Security Verification Standard (MASVS): per OWASP's own MASVS documentation, the MASTG "describes the technical processes for verifying the controls listed in the MASVS." Where MASVS states *what* a mobile app should do to be considered secure, the MASTG supplies the *how* — concrete, per-platform test procedures for Android and iOS, organized around the same eight control categories (MASVS-STORAGE, -CRYPTO, -AUTH, -NETWORK, -PLATFORM, -CODE, -RESILIENCE, -PRIVACY). The project's own GitHub description states it "is a comprehensive manual for mobile app security testing and reverse engineering."
What changed in MASTG v2.0.0, and how is the guide structured now?
MASTG v2.0.0, released 30 June 2026 per the OWASP/mastg GitHub release page, marks the completion of a multi-year refactor that broke the guide apart from large per-topic markdown pages into individually referenceable, machine-readable components: discrete tests (each with its own MASTG-TEST-#### identifier), techniques, tools, best practices, demos, and knowledge-base articles, all cross-linked to the MASVS control and MASWE weakness they address. At the v2.0.0 release, the GitHub release notes put the totals at 193 tests, 152 demos, 167 techniques, 140 knowledge articles, 72 best practices, and 135 tools — counts that are a snapshot of that specific release and will have grown since, since the guide is under continuous development.
What is MASWE, and how does it relate to MASVS and MASTG?
MASWE (Mobile Application Security Weakness Enumeration) is a separate OWASP catalog of specific, named mobile-app weaknesses (for example MASWE-0001, "Sensitive Data Stored Unencrypted in Private Storage"), each mapped to the MASVS control it violates and to the MASTG test(s) that can detect it. OWASP announced MASWE v1.0.0 in a MAS-project blog post dated 17 August 2026, titled "MASWE v1.0.0 is Here." In effect, MASWE sits in the middle of the other two standards: MASVS states the control, MASWE names the concrete way that control can fail, and MASTG supplies the test procedure that proves whether the failure is present.
Who maintains the MASTG, and is it free to use?
The MASTG is maintained by the OWASP Mobile Application Security (MAS) project — the same project that maintains MASVS and MASWE — as an open OWASP Flagship project on GitHub (OWASP/mastg). It is fully free to read online at mas.owasp.org, and OWASP licenses the repository under Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0), matching the license OWASP applies to the MASVS. As with MASVS, OWASP does not certify vendors, testers, or tools against the MASTG; a third party's claim of "MASTG-based" testing is that party's own characterization, not an OWASP endorsement.
Sources
- OWASP Mobile Application Security: MASTG landing page
- OWASP MASVS: Using the MASVS — defines MASTG's role relative to MASVS
- OWASP MASTG: Overview — scope and structure
- GitHub: OWASP/mastg — repository description and license (CC BY-SA 4.0)
- GitHub: OWASP/mastg release v2.0.0 (30 June 2026) — test/demo/technique/knowledge-article counts
- OWASP Mobile Application Security: MASWE landing page
- OWASP MAS blog: "MASWE v1.0.0 is Here" (17 August 2026)
- OWASP MAS: MAS Testing Profiles (L1, L2, R, P, EUDIW) and attacker models
- OWASP MAS: Using MAS Profiles (selecting and tailoring profiles)
- OWASP MASTG: tests index with platform, profile and status tags (read 2026-11-09)
- OWASP MASTG-TEST-0235: Android App Configurations Allowing Cleartext Traffic
- OWASP MASVS-NETWORK-1 control page
- OWASP MASWE-0026: Network Traffic Not Encrypted
About this guide
SoftWSP Firmware Security Desk
Covers the Firmware vulnerabilities & fixes cluster: 131 guides.

In this cluster
Questions about this page?
