InfoQ Homepage Application Security Content on InfoQ
-
GitHub Expands Secret Scanning with General Availability of MCP Server Integration
GitHub has announced the general availability of secret scanning support through its MCP Server, extending automated credential detection and remediation capabilities into AI-assisted and agent-driven development workflows.
-
Attacker Bought 30 WordPress Plugins on Flippa and Backdoored All of Them
An attacker purchased 30+ WordPress plugins on Flippa for six figures, planted a PHP deserialization backdoor in the first commit, and waited eight months before activating it across 400,000 installations. The attack used Ethereum smart contracts to resolve C2. WordPress.org has no mechanism for reviewing plugin ownership transfers, a gap that npm and PyPI addressed years ago.
-
GitHub Enhances CodeQL with Declarative Security Modeling for Faster, More Flexible Analysis
GitHub has introduced a significant update to its CodeQL engine, enabling developers to define custom sanitizers and validators directly through "models-as-data," a move that simplifies how teams extend security analysis across their codebases.
-
Meta's Approach to Migrating their Systems to Post-Quantum Cryptography
Meta has already begun preparing for the threats posed by quantum computing and migrating its systems to post-quantum cryptography, a complex process that will take multiple years to complete. In a recent article, Meta researchers outline their strategy and share key lessons learned along the way.
-
HashiCorp Vault 2.0 Marks Shift to IBM Lifecycle with New Identity Federation
HashiCorp has released Vault 2.0, moving to the IBM versioning and support model following its acquisition. The update introduces Workload Identity Federation for secret syncing without static credentials, SCIM 2.0 provisioning, and performance gains in the storage engine. It also prioritises identity-based security and certificate automation while removing legacy architectural components.
-
New Rowhammer Attacks on NVIDIA GPUs Enable Full System Takeover
Security researchers have demonstrated a new class of Rowhammer attacks targeting NVIDIA GPUs that can escalate from memory corruption to full system compromise, marking a significant shift in hardware-level security risks.
-
Anthropic Releases Claude Mythos Preview with Cybersecurity Capabilities but Withholds Public Access
Anthropic has introduced Claude Mythos Preview, its most advanced AI model, improving significantly in reasoning, coding, and cybersecurity. Unlike previous releases, it will not be publicly available. Access is limited to a consortium of tech companies through Project Glasswing. Internal tests revealed the model's ability to discover critical security flaws effectively.
-
Axios npm Package Compromised in Supply Chain Attack
On March 31, 2026, two versions of the Axios library were compromised and found to contain a Remote Access Trojan. The malicious packages were published through a hijacked maintainer account. The Axios team is investigating how the breach occurred and has deprecated the affected versions. Security experts emphasize the need for better dependency management.
-
TanStack Start Introduces Import Protection to Enforce Server and Client Boundaries
TanStack Start has introduced a import protection, which aims to prevent server and client code from being mixed in full-stack React applications. This Vite plugin automatically checks imports during development and build processes. It blocks harmful imports by file naming conventions or explicit markers, enhancing security and reducing bugs without requiring additional developer input.
-
Sonatype Launches Guide to Enhance Safety in AI-Assisted Code Generation
Sonatype Guide is a real-time guardrail system that sits between AI coding tools and the open-source ecosystem, ensuring AI-generated code uses safe, valid, and maintainable dependencies.
-
WhatsApp Deploys Rust-Based Media Parser to Block Malware on 3 Billion Devices
WhatsApp has rewritten its media handling library in Rust, replacing 160,000 lines of C++ with 90,000 lines of memory-safe code for 3 billion devices. The rollout, part of a system called Kaleidoscope, uses differential fuzzing to ensure bug-for-bug compatibility. The move mirrors a decade-long industry shift toward memory safety, tracing back to Mozilla's first Rust MP4 parser deployment in 2016.
-
How CyberArk Protects AI Agents with Instruction Detectors and History-Aware Validation
To prevent agents from obeying malicious instructions hidden in external data, all text entering an agent's context must be treated as untrusted, says Niv Rabin, principal software architect at AI-security firm CyberArk. His team developed an approach based on instruction detection and history-aware validation to protect against both malicious input data and context-history poisoning.
-
GitLab 18.8 Marks General Availability of the Duo Agent Platform
GitLab 18.8 brings a number of new features, including GitLab Duo Planner Agent, GitLab Duo Security Analyst Agent, auto-dismiss irrelevant vulnerabilities, and more. With this release, the GitLab Duo Agent Platform, enabling organizations to orchestrate AI agents, reaches general availability.
-
Django Releases Version 6.0 with Built-In Background Tasks and Native CSP Support
Django 6.0 has arrived, offering developers cutting-edge features like a built-in tasks framework, enhanced security with Content Security Policy, and template partials for modular design. This release supports Python 3.12 to 3.14, streamlining web app development while promoting clean, rapid design. Join the excitement as Django modernizes the landscape of web application development!
-
Open-Source Agent Sandbox Enables Secure Deployment of AI Agents on Kubernetes
The Agent Sandbox is an open-source Kubernetes controller that provides a declarative API for managing a single, stateful pod with stable identity and persistent storage. It is particularly well suited for creating isolated environments to execute untrusted, LLM-generated code, as well as for running other stateful workloads.