Cyber News Bytes: What’s Happening in Cybersecurity This Week

Another fast week, and the throughline is AI: quietly reshaping how bugs get found, fixed, and exploited, while two of the oldest mistakes in the book still swing doors wide open.

If watching the field move this fast makes you want to be in it, this is exactly the moment my Cyber Interview Prep Course was built for.

It walks you through the technical rounds, the behavioral curveballs, and the "tell me about a recent breach" questions hiring managers love, so you can speak to stories like the ones below with real confidence and land the role.

Read on, then go get ready.

1. Google's AI helped Chrome fix over 1,000 security bugs

Google's Chrome team revealed that AI tools now do a huge share of the browser's security work, contributing to more than 1,000 bug fixes across two recent releases. The standout find was a sandbox escape that had been lurking undetected in Chrome's code for over 13 years.

Google's agents, including Big Sleep and CodeMender, run continuously and blocked more than 20 vulnerabilities from reaching production in a single month. To shrink the window attackers get after a fix goes public, Chrome is now piloting two security releases per week.

Why it matters: This is the optimistic flip side of the AI security story, defenders using the same tech attackers are racing to weaponize. If you are studying for cyber roles, "AI-assisted vulnerability research" is quickly becoming a phrase you want to understand and speak to. Practically, keep Chrome set to auto-update, because faster patches only help if you actually install them.

Read more at BleepingComputer

2. Cisco firewall zero-day exploited through a hard-coded password

Cisco warned that a zero-day in its Secure Firewall Management Center, the console many companies use to run their firewalls, was actively exploited in attacks. The flaw comes from static credentials built into the software, letting an unauthenticated remote attacker log in using that hard-coded account.

The access starts low-privilege, but Cisco warns it can be combined with other flaws to elevate privileges. CISA added it to its Known Exploited Vulnerabilities catalog and ordered federal agencies to fix it by August 1.

Why it matters: Hard-coded credentials are one of the oldest mistakes in the book, and they still show up in products from the biggest vendors. Anything that manages security tooling, like a firewall console, is a prize target, because it hands attackers a map of the whole network. If you run Cisco gear, patch now and keep that management interface off the public internet.

Read more at BleepingComputer

3. A critical TeamCity flaw lets attackers run code without logging in

JetBrains patched a critical vulnerability in TeamCity, the popular tool teams use to build and ship software. Rated 9.8 out of 10, the bug lets an unauthenticated attacker bypass authentication and run operating system commands on the server.

The weakness sits in the agent polling protocol and affects every on-premises version of TeamCity. Cloud customers are safe, but self-hosted users need to update to the fixed versions right away.

Why it matters: Build servers sit at the heart of the software supply chain, so a takeover there can poison everything a team ships. This is exactly the kind of "one unpatched internal tool" story that turns into a major breach months later. Supply chain security and CI/CD hardening are booming interview topics, and this is a clean example to have ready.

Read more at SecurityWeek

4. A researcher used AI to build a Linux root exploit

A researcher at STAR Labs published a Linux kernel exploit that turns an ordinary local user into root, and said AI helped him find the bug and speed up building the attack. The flaw, tracked as CVE-2026-53264, is a use-after-free race in the kernel's network traffic-control code.

AI reportedly optimized the hardest part, cutting the time needed to win the race from more than 15 minutes down to about 5 seconds. The reassuring part is that it is local-only, needs specific conditions, and has already been patched upstream.

Why it matters: Pair this with the Chrome story and you see the whole picture, AI is speeding up both the people fixing bugs and the people exploiting them. The barrier to writing advanced exploits is dropping, which raises the stakes for patching quickly. Make sure your Linux systems run a distribution kernel that carries the fix, not just a matching version number.

Read more at The Hacker News

The theme this week is that AI is quietly rewriting both sides of security at once, helping Google squash a 13-year-old Chrome bug while helping a researcher build a Linux root exploit in record time.

At the same time, the oldest mistakes, a hard-coded password and an unpatched build server, are still the ones swinging doors wide open.

The tools keep getting smarter, but the fundamentals are still what get you hired, and what keep you safe.

Stay sharp, Sandra