SPA: Securing Persistent LLM Agents across Queries with Plan-First Information-Flow Control

Large language model LLM agents increasingly operate over untrusted webpages, documents, tools, and persistent states while exercising authority over security-sensitive resources. Existing defenses typically protect either planning or individual tool interactions, but persistent agents face a...

How to Gain Control of AI Agents and Non-Human Identities

We hear this a lot: “We’ve got hundreds of service accounts and AI agents running in the background. We didn’t create most of them. We don’t know who owns them. How are we supposed to secure them?” Every enterprise today runs on more than users. Behind the scenes, thousands of non-human identitie...

A New Maturity Model for Browser Security: Closing the Last-Mile Risk

Despite years of investment in Zero Trust, SSE, and endpoint protection, many enterprises are still leaving one critical layer exposed: the browser. It’s where 85% of modern work now happens. It’s also where copy/paste actions, unsanctioned GenAI usage, rogue extensions, and personal devices crea...

Information Disclosure

github.com/containerd/imgcrypt is vulnerable to information disclosure. Remote unauthenticated attackers are able to gain access to view an image without providing keys if that image had previously been decrypted. This is possible because only the first architecture in an list is being validated...