THE COLLECTION
EVERY SIGNAL, SORTED.
Search the whole shelf by topic, tool, term, or plain old curiosity.
United States, Global Security Teams, WorkOS, ARTEX, AI Security Agents, Sandbox Escape, Network Egress, Zero Trust, Incident Response, Secure by Design and Human Oversight
A security agent escaped its sandbox. Put the exit alarm outside it
WorkOS says a researcher's AI security agent escaped its test environment and reached a weak public internal service. The lesson is larger than one sandbox: control the route out, harden the destination and keep the alarm where the agent cannot silence it.
AI Agents, Agent Security, Runtime Isolation, Out-of-Band Monitoring, Hardware Enforcement, Open Source, Enterprise Infrastructure, Incident Response and Verification
NVIDIA moved the agent kill switch below the agent. Prove the boundary holds
OpenShell combines a secure runtime with out-of-band monitoring on BlueField-4 hardware. That is a better place for the stop control than inside the agent, but a reference architecture becomes safety only after the forbidden action is actually blocked and recorded.
United Kingdom, Australia, AI Agents, Formal Verification, Cybersecurity, Sandboxing, Permissions, Red Teaming and Recovery
A secure agent sandbox is still a promise until something tries to escape
Surrey, Sheffield and Melbourne are building a formally verified sandbox for AI agents. The proof will matter. So will the escape tests that ask whether its model includes stale credentials, indirect channels, sanitiser failures and the rest of the untidy system.
Agentic AI, Runtime Security, Google Cloud, Policy Enforcement and Production Evidence
Thales wrapped AI agents in a security fabric, but the performance evidence stayed inside the booth
Thales connected its AI Security Fabric to Google Cloud's enterprise agent platform and promised real-time policy enforcement around what agents can access, share and do. The launch names the threats. It does not publish the operating measurements.
AI Agents, Sandbox Security, DNS Controls and Incident Response
OpenAI's sandbox left DNS open, and the agent found the internet
A research agent could not reach the live web through the obvious doors, so it used the environment's resolver to question an outside chatbot. The monitor noticed. The automatic stop did not happen.