Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

At JFrog, our Security Research team continuously monitors and analyzes newly disclosed CVEs across the open-source ecosystem. Throughout our research, we have repeatedly observed cases where the assigned severity score does not accurately reflect a vulnerability’s real-world impact or exploitability. In fact, during 2025, JFrog researchers reassessed NVD critical-severity vulnerabilities and concluded that 96% warranted a lower severity rating.

As organizations aggressively shift from static Large Language Model (LLM) chatbots to fully dynamic, autonomous AI agents (e.g. systems designed to plan workflows, call APIs, write runtime code, and modify enterprise databases), traditional compliance and governance frameworks are hitting a breaking point. A landmark press release from Gartner highlights a critical systemic risk: treating AI agent governance as a monolithic, one-size-fits-all policy guarantees project failure.

There’s a category of security risk that most organizations aren’t ready for. It doesn’t live in your code repository, your CI pipeline, or your developer laptops. It lives in your runtime, in the autonomous AI agents already running in your environment, extending their own capabilities, and making decisions that no human explicitly approved. This is the challenge JFrog set out to address with our integration with NanoCo AI and their open-source agent framework, NanoClaw.

JFrog Security Research recently discovered and disclosed a critical vulnerability in FFmpeg, the world’s most widely deployed media processing framework. The discovered vulnerability, which we’ve named PixelSmash, is CVE-2026-8461 – a heap out-of-bounds write in the MagicYUV decoder (CVSS 8.8 High). We escalated this vulnerability from a simple crash all the way to reliable remote code execution – all it takes is processing a single malicious media file.

It’s official. Gartner just published the very first Gartner Magic Quadrant for Software Supply Chain Security, and JFrog has been recognized as a Leader, placing highest for Ability to Execute among all the vendors included. For an inaugural report in a category this important, that placement means a great deal to us, and we don’t take it lightly.

For years, installing an npm package has meant trusting that every package in the dependency tree will behave as expected. Whether code originated from the npm registry, a Git repository, a remote URL, or an installation script buried deep within a transitive dependency, npm would typically execute or retrieve it automatically during the installation process.

A new CVE drops into a package you depend on. With the JFrog power for Kiro installed, your next move is a single prompt in your IDE, not a tab switch to the JFrog UI and thirty minutes of hand-rolled REST calls. This is what governed agentic development looks like in practice.

Last June, we hosted the first EveryOps Day in Sydney – born from the convergence of DevOps, DevSecOps, and AI/MLOps we were witnessing across every industry in APAC. A year later, with AI’s proliferation across software delivery and security, we took EveryOps Day to Mumbai on May 15, then embarked on the EveryOps Tour: a series of invitation-only executive events across Canberra, Sydney, and Melbourne.

Picture two realities for the same compliance control reaching production. Reality One: Your AppSec team writes a new rule. An engineer uses Claude Code or Cursor to generate the OPA (Open Policy Agent) Rego policy in minutes. They deploy it. It blocks a legitimate release on a missing context variable, and the on-call engineer routes around the gate to ship the code. The AI gave them fast code — but not code they could trust.