Djini AI

AI Mobile & API Penetration Testing

Mon, 01 Jan 0001 00:00:00 +0000

Give every release a mobile pentest.

AI Mobile Pentesting for Banking & Fintech

Mon, 01 Jan 0001 00:00:00 +0000

Put mobile and API pentesting into the release process. Djini operates your app, follows important banking journeys, and gives the mobile and backend teams evidence they can use to fix and retest.

Managed Mobile Pentesting On Demand

Mon, 01 Jan 0001 00:00:00 +0000

Book a scoped assessment of your Android or iOS app and its API backend. Mobile Hacking Lab runs the engagement using Djini, verifies the findings, and gives your team evidence it can act on.

AI Mobile Security Testing for Healthcare

Mon, 01 Jan 0001 00:00:00 +0000

Bring security testing into the delivery of patient and clinician apps. Exercise the running workflow, investigate the API behavior behind it, and keep the evidence usable by the team responsible for the fix.

AI SAST, CI/CD Security Scans & Fix Workflows

Mon, 01 Jan 0001 00:00:00 +0000

Find and fix security issues before the merge.

Custom Local Security Model Fine-Tuning

Mon, 01 Jan 0001 00:00:00 +0000

Turn your organisation’s internal reports, research, and methodologies into a local model that carries your team’s DNA. We help create the dataset and fine-tune a suitable base model of your choice around your collective knowledge.

AI Mobile Pentesting for Ecommerce & Retail

Mon, 01 Jan 0001 00:00:00 +0000

Bring security testing into the pace of mobile commerce. Use Djini to operate checkout and account journeys, review code changes before merge, and give QA agents the devices they need to run.

iOS & Android Device Lab for AI Automation

Mon, 01 Jan 0001 00:00:00 +0000

Build mobile automation on a ready device lab.

Agentic AppSec for SaaS & Enterprise Apps

Mon, 01 Jan 0001 00:00:00 +0000

Connect the security review to the teams shipping the product. Use AI SAST inside development, explore AppSec Portal for coordinated release work, and retain research value through the model-improvement roadmap.

Mobile Exploit Research, Native Fuzzing & PoCs

Mon, 01 Jan 0001 00:00:00 +0000

Turn a research lead into a tested exploit path.

Agentic Security Reviews & AppSec Portal

Mon, 01 Jan 0001 00:00:00 +0000

Delegate a security review and keep the work moving.

Mobile Security & Device Automation for Telecom

Mon, 01 Jan 0001 00:00:00 +0000

Test the subscriber experience on the environment that matters. Connect mobile and API investigations with a lab your automation can use, including supported physical handsets.

Mobile Game Pentesting & Native Exploit Research

Mon, 01 Jan 0001 00:00:00 +0000

Combine app and API testing with research below the UI. Use Djini to investigate player journeys, examine native components, and give your own QA agents devices for repeatable execution.

Private & On-Premise Mobile Security Testing

Mon, 01 Jan 0001 00:00:00 +0000

Run the testing capability where your data belongs.

Mobile Vulnerability Research for CERTs & CSIRTs

Mon, 01 Jan 0001 00:00:00 +0000

Give incident response and vulnerability coordination teams a connected mobile research environment. Investigate Android and iOS components, reproduce reported behavior on devices, and develop evidence for disclosure and remediation.

Security Model Fine-Tuning & Research Loops

Mon, 01 Jan 0001 00:00:00 +0000

Turn validated testing into model improvement.

Advanced Mobile Vulnerability Research for Government

Mon, 01 Jan 0001 00:00:00 +0000

Support specialist law enforcement and government teams researching mobile platforms, native components, and complex vulnerability paths. Connect reverse engineering, coverage-guided fuzzing, and device-based validation in one research workflow.

Mobile Bug Bounty Report Verification with AI Agents

Mon, 01 Jan 0001 00:00:00 +0000

Reproduce incoming reports. Give triage the evidence.

Mobile Bug Bounty & Individual Security Researchers

Mon, 01 Jan 0001 00:00:00 +0000

Hunt mobile bugs. Build the proof. Grow your expertise.

Pwn2Own Ireland 2026: Two Phones, Two Successful Exploits, and Third Place

Thu, 15 Oct 2026 07:22:55 +0000

Last week, I traveled to Ireland to compete in Pwn2Own Ireland 2026. I teamed up with Dimitrios and Tenia Valsamara to form Team ZyGoat.

Keeping with our zygote-inspired team name, we decided to focus exclusively on Android smartphones, specifically the Samsung Galaxy S26 and Google Pixel 10.

Despite competing in only two categories, we successfully compromised both devices, earning a total of $125,000 in prize money, 27.5 Master of Pwn points, and third place overall.

Free Djini AI SAST: Now Fixes What It Finds

Wed, 14 Oct 2026 12:41:43 +0000

Our AI source-code scanner now brings the next step into GitHub: automated fixes, issues and pull requests. Connect your Android or iOS repository, run a scan, and review the proposed changes where your team already works.

Create your free account at app.djini.ai with 10 monthly AI Source Code scans.

Save time. Reduce risk.

Security issues can expose customer data, delay releases and damage trust. But reviewing every code change takes time that development and security teams often don’t have.

Djini AI SAST: A Free AI Scanner for Mobile App Source Code

Tue, 29 Sep 2026 09:31:22 +0000

Today we are releasing Djini AI SAST, a new scanner that finds security bugs in mobile app source code.

Here is the simple version. You point it at your code and it reads through the code, finds the weak spots, and tells you the exact line to fix. It works on Android and iOS. There is no app file to build, no emulator to boot, and no phone to plug in. Most scans finish in under a minute.

Hacking Optimus Prime: One Click to RCE on the TECNO Spark 30 Pro

Tue, 25 Aug 2026 01:08:05 +0000

Story Time

In 2024, I bought an officially licensed Transformers Optimus Prime mobile phone, which was manufactured by Tecno.

Then…it sat in my drawer for a year and a half, because I got too busy.

But then, in December 2025 and January 2026, I decided to use our Djini AI tool to run a Pwn2Own-like assessment against the device. This means I needed to use Djini to analyze ALL of the apps on the device.

Pwn2Own 2025 – Part 1: Samsung Members WebView Takeover & Intent Redirection (CVE-2025-21079)

Thu, 13 Aug 2026 16:49:58 +0000


Product	Samsung Members ( com.samsung.android.voc )
Affected versions	Confirmed exploitable on 5.5.00.13
Fixed in	5.5.01.3
Vulnerability type	WebView Takeover and Intent Redirection
CVE	CVE-2025-21079
Attacker model	Remote Attacker – a malicious website must either be browsed to or clicked on
Tested on	Samsung S25

Introduction

This post is Part 1 of our Pwn2Own 2025 series, covering the vulnerability that provided the initial entry point into the Samsung Galaxy S25 exploit chain we demonstrated at Pwn2Own Ireland 2025.

Technical Advisory – Meta Horizon Shell — Automatic Dangerous Permission Grant via Virtual Input Injection

Mon, 10 Aug 2026 08:13:39 +0000

Note to the Reader

This vulnerability was identified during security research supported by Djini.ai’s Deep Scan capabilities.\

We also want to highlight that this exploit could have been used to enter Pwn2Own Ireland 2025. During that event, a valid entry against the Oculus 3S line of devices would have to demonstrate that a third part application could access the External Camera and Microphone hardware, without user consent.

Technical Advisory – Samsung Dressroom (Wallpaper & Style) – Arbitrary File Write at System UID

Thu, 30 Jul 2026 12:41:32 +0000


Product	Wallpaper & Style / Dressroom ( com.samsung.android.app.dressroom )
Affected versions	Confirmed exploitable on 2.8.10.24 and 2.9.50.89
Fixed in	2.9.50.90 – entries are rejected if the entry contains either ../ or ..\
Vulnerability type	Path Traversal (Zip-Slip) on backup restore → arbitrary file write at UID 1000 → WRITE_SECURE_SETTINGS bypass
CWE	CWE-22 — Improper Limitation of a Pathname to a Restricted Directory (“Path Traversal”)
Attacker model	Local Physical Attacker or Malicious Owner – device must have the screen unlocked, a USB drive must be plugged into the device, and the owner manually drives a SmartSwitch restore
Tested on	Samsung S24, Samsung S25, and Samsung S26

Note

This vulnerability was identified during security research supported by Djini.ai’s Deep Scan capabilities.

Technical Advisory – Meta Quest System-Wide DoS via Unprotected Memory

Wed, 17 Jun 2026 09:05:54 +0000


Product	Meta Quest System Resource ( com.oculus.systemresource )
Affected versions	201.0.0.129.1682 and below (the v201 OS train)
Fixed in	vros 204 — the com.oculus.systemresource application is removed entirely from the build
Vulnerability type	Denial of Service via Uncontrolled Resource Allocation
CWE	CWE-770 (Allocation of Resources Without Limits or Throttling); parent CWE-400
Attacker model	Local — any installed third-party app, no dangerous permissions (only normal, auto-granted FOREGROUND_SERVICE + QUERY_ALL_PACKAGES ). One user interaction: launching the attacker app.
Tested on	Meta Quest 3S

Summary

The System Resource application on Meta Quest devices exposed an exported broadcast receiver (SystemResourceReceiver) and an exported service (SystemResourceService) with no permission gate and no caller validation.\

OAuth in Mobile Apps: How Custom Schemes can leak Your Tokens

Tue, 19 May 2026 13:28:16 +0000

OAuth 2.0: The Protocol

OAuth 2.0 solves a delegation problem: how can a third-party application access resources on your behalf without ever learning your password? The answer is a framework (defined in RFC 6749) that issues scoped, time-limited tokens in place of credentials. It is an authorization protocol — it says what an application is allowed to do, not who the user is. Identity (authentication) is a separate concern layered on top via OpenID Connect (OIDC).

Detecting and Exploiting Insecure Deserialization with Djini.AI

Fri, 15 May 2026 09:52:38 +0000

Djini AI

Insecure deserialization is one of the most dangerous vulnerability classes in mobile applications — and one of the hardest to find manually. When an app reconstructs objects from raw data (Intents, backup files, deep links, IPC messages), an attacker who controls that data can inject malicious objects that execute arbitrary code during the reconstruction process. No authentication bypass needed, no SQL injection — just a carefully crafted data stream that the app’s own code turns into a weapon.

Intro to Intent redirections…and how to exploit them

Thu, 14 May 2026 11:01:59 +0000

Android Intents are the core messaging system that allows components to talk to each other — whether that’s within the same app or across different apps on the device. They power pretty much everything, from launching Activities to broadcasting system events. They’re fundamental to how Android works.

Now, a lot of apps accept external input — through deep links, other applications, or IPC mechanisms — and use that input to construct or redirect Intent objects internally. When an attacker can influence where those Intents get routed, the result is a class of vulnerability known as Intent redirection. And it’s one of the most impactful attack primitives in Android security.

Unlimited Mobile Security Testing, Directly From Your AI Agent

Wed, 13 May 2026 10:54:36 +0000

Unlimited Mobile Security Testing, Directly From Your AI Agent

Your AI agent can now connect directly to the Djini AI Agent and run mobile security tests and tasks in parallel, with no user interaction and no need to open the Djini UI.

Djini publishes an Agent-To-Agent (A2A) compatible endpoint, so any agent that supports the protocol, Claude Code, OpenCode, and others, can connect to Djini – delegate tasks, and receive results automatically.

1-Click ATO: A studycase of common Android apps misconfigurations.

Fri, 24 Apr 2026 00:00:41 +0000

Summary

TL;DR

Initial Analysis

Deeplinks and Their Final Destination

Scheme Routing

The Handler Logic

Automated API Security Testing for Mobile Backends

Thu, 23 Apr 2026 08:33:34 +0000

Description

Your mobile apps talk to backend APIs. Those APIs handle login credentials, personal data, payment information, and business logic. A vulnerability in any of them can mean a data breach, regulatory fines, or loss of customer trust.

The Problem: Three Surfaces, Three Setups, Never Enough Coverage

Most organizations treat mobile security as three separate problems. Testing the Android app requires one set of tools, emulators, and expertise. Testing the iOS app requires another, with different devices, different toolchains, and different skill sets. And testing the backend APIs? That’s a third effort entirely, often owned by a different team with its own scanners, its own test environments, and its own backlog.

Unauthorized Device Command and Control

Sun, 01 Mar 2026 13:08:49 +0000

Summary

The issue was caused by improper access control and missing input validation in the exported activity. The activity was intended to handle queries, but it accepted Intent data from any application without verifying the caller or enforcing permission checks. The supplied extras were forwarded directly into the AI processing pipeline and mapped to device control commands handled by a privileged backend component. Because no caller validation, permission enforcement, or input sanitization was implemented, attacker-controlled input was able to reach system-level APIs and execute sensitive operations.

Intro to Android WebViews and deep links…and how to exploit them

Wed, 28 Jan 2026 07:42:31 +0000

Android WebView is a system component that allows applications to render web content directly inside a native app, and it is one of the most widely used building blocks in the Android ecosystem.

At the same time, deep links have become the primary way apps expose internal functionality to the outside world. When these two components intersect — when a deep link controls what a WebView loads — the result is one of the most common and dangerous classes of mobile vulnerabilities seen in real-world applications.

From WebView to Remote Code Injection

Thu, 22 Jan 2026 14:13:37 +0000

Summary

TL;DR

Initial Analysis

The JSInterface

Checks whether the message is indeed handled by this handler (checks the action attribute).
If yes, it handles it accordingly and gets the args from the params attribute of the message.

Agents

Mon, 01 Jan 0001 00:00:00 +0000

Contact us

Mon, 01 Jan 0001 00:00:00 +0000

Platform

Mon, 01 Jan 0001 00:00:00 +0000

Why Djini

Mon, 01 Jan 0001 00:00:00 +0000