Five Minutes and 88 Kopecks: AI Neural Networks Can Now Disassemble, Modify and Reassemble Android Apps at Minimal Cost
Researchers at Positive Technologies have demonstrated that modern large language models can disassemble, modify, reassemble, and maintain the functionality of Android applications in as little as five minutes for a cost of only 88 kopecks. The findings reveal how dramatically artificial intelligence has lowered the barriers for creating altered copies of mobile apps that could be used for malicious purposes.
The experiment was conducted in an isolated laboratory environment using a sample of 90 Android applications from various categories. Specialists avoided adding any malicious functions, working with real users, or accessing user data. Instead, they introduced neutral code changes and verified whether the applications would still launch successfully after repackaging. This methodology allowed the team to measure how easily neural networks can automate the creation of modified application copies.
Performance of Different AI Models
Closed commercial models successfully completed the task in an average of 84 percent of attempts. Models with open weights achieved a lower but still significant success rate of 61 percent. On average, the process required 14 interaction steps with the model for each application. Depending on the chosen model, the entire workflow took between 5 minutes 38 seconds and 9 minutes 9 seconds. The cost of a successful modification ranged from 88 kopecks to 40 rubles 89 kopecks. Researchers estimate that a budget of just several thousand rubles would be sufficient to attempt modifications on approximately one hundred popular Android applications.
Altered APK files can be distributed under the guise of legitimate programs through unofficial app stores, websites, messengers, and catalogs of third-party builds. These modified versions are frequently advertised as improved applications that remove restrictions or add extra features. Services that are unavailable in official stores are especially vulnerable, as users often search for installation files on alternative platforms and risk downloading visually identical but tampered builds.
Implications and Recommendations
Large language models have not created entirely new attack methods; however, they have significantly reduced the cost and complexity of preparing counterfeit applications. Previously, disassembling and repackaging APK files required advanced reverse-engineering skills and manual effort. Today, a substantial portion of these operations can be performed automatically by neural networks.
Developers are advised to protect client-side code against analysis and modification, regularly test applications for resilience to reverse engineering, and monitor the appearance of unofficial APK files. Users should be warned about the risks of installing applications from unverified sources, and protective mechanisms should be integrated during the development phase itself.
Related articles
Agent-Ops 0.4.0 Released: Methodology for Secure Human-AI Collaboration in IT Operations
Sergey Zhitinsky, founder of Git in Sky, has published the public normative candidate for Agent-Ops 0.4.0, an open industry methodology governing how engineers and AI agents jointly handle IT infrastructure tasks. The framework keeps humans firmly in the decision-making loop while using deterministic programs for data collection and approved changes. It addresses risks such as prompt injection through processed data, unverified model outputs, and unclear accountability when AI recommendations lead to incidents. The methodology divides work across eight explicit steps and three separate planes: data, governance, and independent verification performed by a Guardian role. Two additional companies have joined as maintainers following agreements at the IT Elements 2026 conference, turning the project into a multi-organization effort. Contributors are invited to help refine contracts, schemas, and operational scenarios through GitHub and GitVerse.
ProxyKey MCP: Securing API Access for AI Agents Without Exposing Credentials
ProxyKey has released an MCP server that allows AI coding agents such as Claude Code and Cursor to manage API credentials without ever reading the actual secret values. The solution addresses the risk that any key visible to an agent becomes compromised through logging, tracing, or prompt injection. Real provider keys are stored encrypted with AES-256-GCM and never returned by any API endpoint after initial entry. Agents instead receive limited virtual passes that support IP binding, rate limits, TTL, and detailed request logging. A pending-secret workflow lets agents prepare services before the real token exists, with the human entering the secret only through a web panel. The approach deliberately restricts the MCP tool contract so no operation can read or return secret values.
Shadow AI in CI/CD: Why AI Agents Must Be Modeled as Security Threats
A new analysis from the CNCF highlights the growing risks of Shadow AI within continuous integration and continuous deployment pipelines. The report argues that AI agents should be treated as potential threats rather than simple productivity tools. Starting from a developer's laptop and extending to Kubernetes clusters, these agents can introduce unauthorized access paths and data exposure risks. Security teams are urged to incorporate AI agent behavior into formal threat modeling exercises. The discussion emphasizes the need for visibility and control over autonomous AI components operating in production environments.
Detecting Lateral Movement with Neural Networks Trained Solely on Synthetic Data
A researcher generated entire corporate network histories using a 135-line configuration file to create synthetic authentication logs containing lateral movement attacks. Neural networks trained exclusively on these artificial datasets were then evaluated against 1.65 billion real authentication events from Los Alamos National Laboratory, including 749 red team events across 301 compromised machines. The best ensemble of six models flagged 3.6 million hourly machine windows and placed 16 genuine attacks among the top 23 highest-scoring entries, producing only seven false positives. In comparison, a simple threshold counter required 161,000 false alarms to reach the same detection level. The approach also demonstrated an iterative feedback loop where detector errors directly informed refinements to the synthetic world generator. The work shows that synthetic data can reach AUC performance comparable to models trained on real labeled attacks while providing full control over the underlying attack definitions.