CrowdStrike, working with Google and the Shadowserver Foundation, has successfully disrupted the GlassWorm botnet that has been targeting open source software developers for over six months. The sophisticated malware used multiple resilient command-and-control channels including the Solana blockchain, Google Calendar, BitTorrent, and traditional servers, and was designed to steal developer credentials and cryptocurrency while hiding itself using invisible Unicode characters in code editors. The takedown is particularly significant because GlassWorm represents a growing threat where attackers target developers themselves rather than just software products, creating serious supply chain security risks for any organization that consumes software built in compromised development environments.

Researchers at Adversa AI have discovered "SymJack," a new attack method that exploits AI coding agents to inject malicious code into software development pipelines by hijacking symlinks in project files. The attack works by disguising a malicious symlink as an innocuous file that, when approved by an unsuspecting developer, secretly registers a malicious server that can steal credentials and access tokens. Testing across five major AI coding agents—including Claude Code, GitHub Copilot, and Cursor—found all were vulnerable, though most vendors rejected the report, with only Anthropic later quietly adding protections to display real symlink destinations before requiring approval.

Stolen credentials have become one of the most effective attack methods in cybersecurity because when attackers use legitimate login information, they essentially appear as authorized users, making detection extremely difficult. The problem spans both human credentials like passwords and biometrics, as well as non-human credentials like API keys and session tokens, with companies potentially managing hundreds of thousands of active tokens. While tools like dark web monitoring and services like Have I Been Pwned can help detect compromised credentials, experts say preventing theft is nearly impossible due to sophisticated phishing, social engineering attacks, and infostealer malware that continuously scrape credentials from infected systems.

Canadian identity security company Lastwall has raised 11.5 million dollars in a Series A extension to expand its quantum-resilient identity platform across North America. The company provides passwordless authentication and identity management for defense and government organizations, with its flagship product evaluating 200 contextual signals like keystroke patterns and mouse movements to assign risk scores to login attempts. Lastwall also offers Quantum Shield, a quantum-safe encryption solution that applies NIST-approved post-quantum cryptography algorithms to protect data in transit while maintaining backward compatibility with conventional encryption.

A Romanian hacker has been sentenced to four years and eight months in a US prison after pleading guilty to breaching an Oregon state government network and selling access to it for three thousand dollars in Bitcoin. Forty-five-year-old Catalin Dragomir admitted to hacking and selling access to at least ten organizations, causing losses exceeding two hundred fifty thousand dollars, though he claimed he was working for another hacker rather than masterminding the scheme himself. Dragomir was arrested in Romania last November and extradited to the United States in January before entering his guilty plea.

London-based cybersecurity startup RevEng.AI has secured 15 million dollars in Series A funding, led by the NATO Innovation Fund, bringing its total funding to 19.5 million dollars. The company uses an AI model called BinNet to analyze compiled software at the binary level, hunting for vulnerabilities and backdoors without requiring access to source code. This capability is becoming increasingly critical as AI coding agents produce more autonomous software that may never be reviewed by humans, giving organizations an independent way to verify closed-source, third-party, and internally developed software before deployment.

SecurityWeek is hosting its third annual AI Risk Summit on August 11th and 12th, 2026, in Half Moon Bay, California, bringing together security leaders, AI researchers, and policymakers to tackle the challenges of AI adoption and security. The conference will cover topics including securing AI deployments, adversarial attacks, deepfakes, governance frameworks, and data protection, with early registration available at a discounted rate through May 31st. The summit runs alongside SecurityWeek's CISO Forum Summer Summit, giving attendees access to both events for expanded networking on cybersecurity and AI risk management.

AI chatbots are being exploited to direct users to malicious websites that install cryptojacking malware, which secretly uses victims' computer resources to mine cryptocurrency. Security researchers have discovered that attackers are manipulating chatbot recommendations to serve these harmful links, representing a new evolution in how AI systems can be weaponized to distribute malware. This development highlights growing concerns about the security vulnerabilities in AI-powered recommendation systems that millions of users rely on daily.

A critical security flaw has been discovered in Gitea, the popular open-source software development platform, that allows attackers to access private container images without any authentication. The vulnerability effectively bypasses security controls, potentially exposing sensitive proprietary code and container configurations to unauthorized users. Organizations using Gitea for container registry services are urged to patch immediately to prevent unauthorized access to their private repositories.

Shadow AI—where employees use unauthorized AI tools—is becoming a major security challenge for companies. IT leaders are struggling to balance the need for security controls with keeping workers productive, as banning these tools outright often just drives usage further underground. The solution involves creating clear policies, offering approved AI alternatives, and implementing monitoring systems that guide rather than restrict employees while protecting sensitive company data.