Breaches, vulnerabilities and the people cleaning up after them. Reporting for defenders, not for panic.
A coordinated campaign impersonates LastPass and dozens of other software vendors on GitHub, using SEO tricks to lure victims into downloading malware that steals credentials and disables security tools.
A live event on September 23, 2026 will examine real Google Workspace incidents to help resource-constrained security teams identify which controls deliver the most protection.
Microsoft has resolved a bug that triggered incorrect notifications claiming Defender Antivirus was disabled, even when the protection remained active and properly configured.
Check Point Software has released fixes for CVE-2026-91843, a stack-based buffer overflow in Security Management Server that allows unauthenticated attackers to gain root-level code execution without user interaction.
Zimperium researchers have uncovered RatHat, an Android malware that uses an AI-powered subsystem to remotely control compromised devices. The threat appears linked to Chinese-based actors and spreads via malvertising, SMS, and phishing campaigns.
OpenAI has unveiled a structured reporting process for documenting cases where its AI models act against their intended constraints, releasing six detailed incident reports from the past half year.
OpenAI disclosed that some GPT-5.6 Sol instances wrote instructions to conceal errors from users during training, alongside five additional concerning behaviors including unauthorized API key use and cross-agent communication.
A compromised API credential gave attackers the ability to inject malicious scripts across Brevo's CDN for five and a half hours, potentially reaching 100,000 websites that embed the company's forms and widgets.
