Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Passwords were invented in the 1960s. Six decades later, we’re still using them to protect everything from email accounts to bank transfers to corporate networks. The problem isn’t just that they’re old technology, it’s that they were never designed for the world we live in now.

Microsoft was the most commonly impersonated brand in phishing attacks during the fourth quarter of 2025, according to researchers at Guardio. Microsoft was followed by Facebook, Roblox, McAfee, Steam, AT&T, Amazon, Google, Yahoo, and Coinbase. “Scammers ramped up brand impersonation attacks throughout Q4 2025, timing their campaigns around when people are busiest online, shopping for deals, renewing subscriptions, or looking for jobs,” Guardio says.

Researchers at RavenMail warn that a major phishing campaign targeted more than 3,000 organizations last month, primarily in the manufacturing industry. The phishing messages posed as legitimate business notifications, such as file access requests or voicemail alerts, and were designed to send users to credential-harvesting login pages. Notably, the campaign abused legitimate Google infrastructure and links to avoid being flagged by security tools.

As per Traceable’s 2025 State of API Security report, only 21% of the >1500 respondents surveyed across the globe showed confidence in detecting attacks at the API layer. Furthermore, only 13% were capable of preventing >50% of API attacks. This is when the API sprawl is still burgeoning. The challenge, thus, is no longer volume but maturity.

The recent disclosure of what security researchers are calling "the most severe AI-driven vulnerability uncovered to date" in ServiceNow's platform serves as a stark reminder: securing agentic AI isn't just about new AI-specific controls; it requires getting the fundamentals right first.

At JFrog, our mission has long been to power the future of software, and we believe that future is undeniably cloud-native. This is why we’ve architected our platform as a container-first, Kubernetes-native SaaS—built for performance at scale on the world’s leading cloud infrastructure. Our deep commitment to cloud excellence has reached a major milestone in our long-standing collaboration with Amazon Web Services (AWS): JFrog has achieved AWS Security Competency status.

A recent stack buffer overflow vulnerability in Redis, assigned CVE-2025-62507, was fixed in version 8.3.2. The issue was published with a high severity rating and assigned a CVSS v3 score of 8.8. According to the official advisory, “a user can run the XACKDEL command with multiple IDs and trigger a stack buffer overflow, which may potentially lead to remote code execution”.

To kick off 2026, I’m proud to share that we’ve released the inaugural edition of the SafeBreach State of the Breach Report. This report has roots going back over 11 years when SafeBreach was originally founded. Even then, our goal was always to empower security leaders to better understand the efficacy of their security programs and make data-driven decisions—no more guessing what to do.

AI systems are moving fast, sometimes faster than the guardrails meant to contain them. In this episode of Security Matters, host David Puner digs into the hidden risks inside modern AI models with Pamela K. Isom, exploring the governance gaps that allow agents to make decisions, recommendations, and even commitments far beyond their intended authority. Isom, former director of AI and technology at the U.S.

As cloud environments sprawl and engineering teams scale, the number of identities you manage has exploded. It’s no longer just employees and contractors; CI/CD pipelines, service accounts, API tokens, and AI-powered agents are all asking for access to production systems and sensitive data. It’s no shock that identity has become a top-line priority for security and platform leaders.