PPhying中文
Curated security research, vulnerabilities, advisories and tools for practitioners.

DeepsecBench: evaluating model performance in finding cybersecurity vulnerabilities

vercel.com | research | #ai-security | #cloud-security | #vulnerability-research | #web-security | #vercel
📄 Raw .md⬇️ Download .md🔗 Original ↗

Summary

Today we're releasing DeepsecBench, a benchmark that evaluates how well different models find security vulnerabilities in application code.
Published
Collected

Related coverage

research · vercel.com
Security boundaries in agentic architectures
A framework for drawing security boundaries in agentic architectures. Most agents run with zero isolation between the agent and the code it generates. Learn where to draw the boundaries, from secret injection to full application sandboxing.
vulnerability · vercel.com
Reproducing, disclosing, and fixing the libheif vulnerability with Hacktron and the maintainers
How Vercel and Hacktron coordinated a fix for a remote code execution vulnerability that reached Next.js image optimization through sharp, libvips, and libheif.
vulnerability · vercel.com
The Vercel OSS Bug Bounty program is now available
Vercel is opening its open source software bug bounty program to the public for researchers find vulnerabilities and make OSS safer
blog · vercel.com
The three types of AI bot traffic and how to handle them
Bots account for over 20% of all web traffic. A quarter of that is AI crawlers alone. But not all bots are bad. Some are key to how your site gets discovered. Learn how AI bots shape visibility and what happens if you block them.
vulnerability · vercel.com
Postmortem on Next.js Middleware bypass
Last week, we published CVE-2025-29927 and patched a critical severity vulnerability in Next.js. Here’s our post-incident analysis and next steps.
research · vercel.com
Everything hackable will get hacked
Open-weight models that allow perform offensive research present an emerging cybersecurity challenge. But you do not need to wait for specialized models to begin defensive cybersecurity work.
Back