Key takeaways: Cloud security and AI security stop being two things For the last couple of years, a lot of the security industry has been quietly insisting that AI security has to be its own thing. Keep tools, teams, budgets, and dashboards separate or else… Well, we disagree. We listened to our customers, and what…

Key takeaways: The CISO squeeze in 2026 If you’re a CISO right now, you already know the feeling. The board wants an AI strategy immediately. Engineering is already shipping copilots, agents, and Gen-AI workflows into production. Teams are connecting models to APIs, internal systems, SaaS platforms, and cloud infrastructure faster than governance can keep up.…

Key Takeaways What "AI visibility gap" actually means I've spent most of my career running security at companies that were three steps behind their own infrastructure. No judgement because in many ways that's the job… you catch up, the platform team ships something new, you catch up again. You could consider it a treadmill to…

Key Takeaways Closing the gap, in practice The visibility gap post made the strategic case: cloud security tools were built for static infrastructure, AI workloads are runtime workloads, and the snapshot model can't see them. The three capabilities that actually close it are AI-Inventory, AI-BOM, and AI-NHI, and I'll walk through each one in order:…

Key Takeaways Listen in on most security teams talking about prompt injection in 2026 and you'll hear something like, "it's a model problem, we have guardrails, we're fine." Three months later an attacker is exfiltrating customer records through that same model, and the SOC has no clue why their CDR didn't fire. Prompt injection is…

Key Takeaways A silent data bleed looks like this. Somewhere in your cloud, an EC2 workload is making an HTTPS call to api.deepseek.com or api.openai.com or api.anthropic.com. The TLS handshake is identical to every other vendor API call in your environment, the destination is fronted by a CDN your monitoring already trusts, the payload is…

Key Takeaways Allow me to let you in on a methodology secret Most security teams I talk to now have budget allocated for AI security testing in 2026. Most are spending it the same way they spent security budgets before AI existed: I’ve done the same thing myself more than once. It’s the path of…

Key Takeaways The CISO bottleneck in 2026 If you're a CISO in 2026, you're probably the slowest part of your company's AI strategy. Not because you're slow, but because the rest of the org can stand up new AI features in days, while the security review cycle is still measured in weeks. Product is shipping…

Key Takeaways Agentic cloud security is everywhere. What's actually real? Every cloud security platform shipped AI agents this year, but the differentiator isn't the agents. It's the data architecture they reason on. Investigation agents, remediation agents, triage agents, copilots, analyst agents, workforce agents. Rest assured, most of these are real systems. But some are LLM…

Executive Summary A new wave of the Mini Shai-Hulud campaign compromised dozens of official @tanstack/* npm packages by abusing CI/CD publishing workflows and trusted npm release mechanisms. Unlike traditional dependency malware focused only on downstream execution, this operation behaves as a self-propagating supply chain worm designed to continuously spread across repositories, developer environments, and CI/CD…

Security teams already have enough findings to sort through. Vulnerabilities, misconfigurations, and alerts pile up every day, but only a limited number create real risk in production. The harder problem is knowing which issues are exposed, active, and worth fixing first. The Upwind MCP Server connects MCP-compatible clients to selected Upwind context and capabilities, helping…

Executive Summary Dirty Frag is a new Linux kernel local privilege escalation that combines two kernel bugs - one in the IPsec subsystem and one in RxRPC, giving any unprivileged local user a root shell on every major distribution. The exploit is reliable and lasts until a reboot or cache fault Public PoC code has…