Key Takeaways: Agentic AI security is an architecture problem, not a policy problem. Most organizations have adopted AI agents in the form of coding assistants, autonomous workflow tools, internal chatbots connected to production systems, but without establishing the foundational security frameworks those systems require. The adoption pressure is real. Telling your engineering team to stop…

We're excited to share that Frost & Sullivan has recognized Upwind as a Strong Visionary Leader in the Frost Radar™: Cloud/Application Runtime Security, 2026. This recognition highlights the company's innovation, growth, and leadership in the emerging Cloud-Native Application Detection and Response (CNADR) market. For us, the recognition is meaningful not simply because of where Upwind…

Every API has a different risk profile. An internal billing endpoint and a public-facing authorization endpoint don't fail the same way. They don't get attacked the same way either. A generic ruleset can't account for that. Custom rules can, and now those rules can see sensitive data too. This new release brings two things together…

Kubernetes environments move fast. Workloads appear and disappear, container images change continuously, services are exposed, permissions evolve, and development teams deploy updates throughout the day. But most cloud security platforms force practitioners to investigate Kubernetes risk through interfaces designed for the broader cloud, leaving teams to manually filter the noise before they can begin investigating…

Frontier AI models combined with a rampant rate of new critical vulnerabilities mean speed and context are everything. When a critical production service starts behaving suspiciously, every second spent jumping between different tools and dashboards is a second lost to a potential attacker. At Upwind, we are excited to introduce a game-changer for security teams:…

Executive Summary Upwind identified a critical supply chain compromise across five npm packages in the @asyncapi scope, published on July 14, 2026 via two separate branch compromises in two GitHub repositories. The attacker never touched an npm token. They abused each project's own CI pipeline through GitHub Actions OIDC to publish the malicious packages. The…

Let this sink in: finding security vulnerabilities was never the bottleneck, vulnerability prioritization was. AI-scale discovery is about to make that impossible to ignore, and the teams that see it coming will pull ahead of the rest. Here's what kept me up this week. Anthropic ran a frontier model against some of the world's most…

Software supply chain risk doesn't stop at the container boundary. Most organizations still run a meaningful share of production workloads on virtual machines across legacy services, data pipelines, and infrastructure that was never containerized. The Upwind Platform creates SBOMs at runtime, delivering greater accuracy than build-time tools by continuously monitoring your live environment. Format adds…

If you follow AI at all, you know the leaderboards. Every few weeks a model takes the top spot, and we all check where our favorite landed. But a leaderboard only tells you who's ahead, and it stays quiet about where any of those models still come up short. Which, conveniently, is the part that…

Focus Mode is now available in the Upwind platform, giving security teams a faster, more focused way to work. Instead of navigating across the platform, you can switch to Focus Mode to slice and dice the Upwind platform by Vulnerability Management, Cloud Security Posture, Attack Surface Management, Administration, or Threats, and starting next week, AI…

We’re excited to introduce Early Advisories as part of the Upwind platform. Powered by Upwind's security research team, Early Advisories notify customers about emerging threats, including zero-days and supply chain attacks, before they receive a CVE identifier. Early Advisories are published directly into the Vulnerabilities module alongside CVE-based findings and automatically correlated with your live…

Key Takeaways When Anthropic published its piece on recursive self-improvement, the line that traveled was the efficiency one. A 100-person company doing the work of a much larger one, because each person sits atop a pyramid of agents. It's a striking image, and it's probably right. But there's a second diagram hiding inside the first…