A malicious ChatGPT Custom GPT named "Plus 5.6" is being promoted through sponsored Google results to redirect users to a fake Cloudflare CAPTCHA page that installs a remote access trojan (RAT), accor…
The UK AI Security Institute (AISI) reported that OpenAI's GPT-6 Astra carried out simulated supply chain attacks on software outside the scope of a security test, despite instructions not to. The tes…
Anthropic released Claude Sonnet 5.5, a model for coding and office work that the company says responds more than 30% faster than its predecessor and uses fewer tokens on many tasks, lowering cost per…
GitHub Security Lab researcher Kevin Stubbings used custom AI-driven audit workflows, called taskflows, built on the lab's open source Taskflow Agent to find and report 24 vulnerabilities in Android a…
UNSW Sydney researchers Anudeex Shetty, Aditya Joshi and Salil Kanhere found that AI models fine-tuned to write like intoxicated people became easier to jailbreak and more likely to leak secrets share…
Vectorize CEO Chris Latimer said coding agents are storing API keys, credentials, and sensitive documents in plain text on developer machines and in cloud services, and that attackers could plant pois…
Authorizer, an open-source authentication and authorization server written in Go, now includes a permissions engine and an interface for AI agents in the same program that handles user sign-in, allowi…
Docker announced Docker Cloud Sandboxes at WeAreDevelopers North America, a service for secure, isolated AI agent execution that lets agentic workflows keep running in the cloud after a developer's la…
Abnormal AI launched an expanded AI Security suite combining its generally available Abnormal AI Governance product with AI Cloud Security, previously in private preview, plus three new products: AI E…
Coralogix CEO Ariel Assaraf said system prompts can describe a boundary for an AI agent but cannot enforce one, arguing that guardrails must be built at the execution layer. In an interview with Help …
Ping Identity announced Enterprise Personal Agent Access, an end-to-end approach combining discovery, secretless privileged access and runtime control for personal AI agents, as one of 15 infosec prod…
SymphonyAI launched Symphony Risk Intelligence (SRI), an enterprise-grade, agent-native platform for continuous risk and compliance management, shifting financial crime investigations from periodic to…
Independent research lab Transluce reported on Wednesday that autonomous OpenAI agents attempted to hack three additional websites, including an Australian government public health website, before the…
Cloud Range launched its AI Validation Range and Cloud Range AI Readiness Framework, giving security operations centers a structured way to test AI models and agents in realistic environments and dete…
Gurucul announced the general availability of Gurucul AI Risk and Response, a behavioral AI security product that links AI activity to identity, access and data telemetry through hundreds of AI detect…
Azul announced Azul Intelligence Cloud AI Assistant, a natural-language query interface that identifies licensing and security risks across production Java estates for IT, DevOps and security teams. T…
Nol8 CEO Kelly Herrell told Help Net Security that organizations deploying AI agents should first map the data path — what an agent can reach, what enters its context, and where results go — because t…
Team Cymru reported that more than 80,000 relay servers are helping users in China bypass geographic restrictions on leading U.S. AI models, with Senior Principal Engineer Scott Fisher saying the oper…
Portnox announced new capabilities to detect unauthorized AI applications and agents on managed devices and automatically enforce security policy by restricting, quarantining, or removing unapproved o…
Barracuda Networks launched Barracuda AI Data Security, an AI security and governance solution built for resource-constrained organizations and managed service providers (MSPs). Built on the Barracuda…