← Back to Blog
Security notice

Next.js Is Using AI to Find Its Own Security Holes: What Agencies Running Next.js Should Know

The Next.js team is using AI to audit the framework itself. More advisories are coming intentionally. Here's what that means for agencies in production.

Next.jsNovember 18, 20265 min readBy Joseph Rajewski
Next.js Is Using AI to Find Its Own Security Holes: What Agencies Running Next.js Should Know

If you've noticed more Next.js security advisories showing up in your feed lately, that's not an accident, and it's not a sign that the framework is falling apart. The Next.js team published a post this week explaining exactly why advisory volume is rising and what they're doing about it. For agencies running Next.js for clients, the distinction between "more bugs found" and "more bugs exist" matters a lot, and we want to unpack it.

What the Next.js team published

The post, titled "How we're using AI to make Next.js more secure," lays out why you are seeing a higher frequency of security advisories coming from the project. The short version: the team is deliberately using AI tooling to audit the framework's own codebase, and that audit is finding issues that manual review had not caught. The blurb from the post makes this explicit: more advisories are the expected output of a more thorough process, not evidence of a deteriorating security posture.

The post does not describe a single critical zero-day or active exploit. What it describes is a systematic, proactive security program being applied to a widely deployed framework. That framing is important. The team is surfacing issues before attackers do, which is the right order of operations.

What we don't have from the blurb alone is a specific count of vulnerabilities found, their severity levels, or the Next.js version range affected by any individual fix. The advisory linked in the post is the place to confirm those specifics. If you are running Next.js in production, that post is worth reading in full, not just bookmarking.

Our take

We've been running Next.js on client sites for a few years now, and our honest reaction to this news is that it's genuinely good, even if it's going to create more short-term work for agencies.

Here's the uncomfortable math: if you're responsible for 5 or 10 Next.js sites, and the advisory cadence doubles because the framework team is now running AI-assisted audits, your patching workload also roughly doubles. That's a real operational cost. It doesn't mean the framework is worse. It means the people maintaining it are doing a better job of finding problems before attackers do. We'd rather have that tradeoff than the alternative, which is a quieter advisory feed masking unfound vulnerabilities.

The more important point for agencies is what this signals about where the industry is heading. Using AI to audit framework source code is not a Next.js-exclusive idea. It's a technique the Vercel team is applying today, and it's one that will spread to other major frameworks and CMS platforms over the next 12-24 months. We already wrote about the Drupal AI Security Initiative finding 256 advisories in five months using similar approaches. The pattern is consistent: AI-assisted audits find more, faster, and the short-term noise is worth the long-term improvement in baseline security.

For clients, the conversation is also shifting. "Why did you have to patch Next.js again this month?" is a question we're already fielding, and the honest answer is: because the framework team is doing its job. That's a much better answer than "we don't know why there's a vulnerability."

One practical concern worth naming: not all agencies have Next.js sites on a managed update workflow. If you're running a Next.js deployment on a custom server or a container setup where someone has to manually trigger a framework update, a higher advisory cadence puts real pressure on that process. This is a good moment to audit whether your Next.js client sites have a defined patch path, not just a deployment pipeline.

What to do right now

Read the full post. The blurb is enough context to understand the why, but the Next.js advisory page will tell you whether any specific version you are running is affected by a current fix. Check it at nextjs.org/blog/how-were-using-ai-to-make-nextjs-more-secure.

Audit your Next.js version coverage across client sites. Know what version every client site is running. If you don't have that in a spreadsheet or dashboard somewhere, now is the time to build it. Advisory fatigue is real, and the only defense against it is having a clear, low-friction way to know who needs what.

Set up automated dependency alerts. If you're not already running Dependabot or an equivalent tool on every Next.js repository you maintain, turn it on today. When the advisory cadence rises, manual monitoring doesn't scale.

Budget for patching time in retainer agreements. If a client's maintenance retainer doesn't explicitly cover security patching time, this is the conversation to have before an urgent advisory forces it. A rising advisory cadence from a major framework is exactly the scenario those retainer clauses were designed for.

Don't treat silence as safety. A framework with a low advisory count isn't necessarily more secure than one publishing frequently. It may just have a less rigorous audit process. Vercel is demonstrating that proactive auditing is a feature, not a liability. We think that framing is correct.

Originally referenced: How we're using AI to make Next.js more secure on Next.js Blog.

If you're running Next.js in production and want help building a reliable patching workflow or understanding how a specific advisory affects your client sites, get in touch.

#nextjs#security#ai#releases#agency

Need help with your project?

Let's discuss how Digital Pixel can help bring your vision to life.

Get in Touch