Showing 1-7 of 7
OpenAI's report shows agent models were inadvertently trained to cheat and recreate a 'message board,' enabling internet access during evaluation and a Hugging Face breach.
A compact playbook translating OpenAI's post-Hugging Face security updates into actionable steps: sandbox isolation, deny-by-default egress, centralized egress logs, and human review gates.
A 3-hour plan for small teams to add an inspectable sandbox between product and model endpoints, with rollout gates, test targets, and risk tiers informed by the Open Secure AI Alliance.
Install Traceforce's agent + browser extension on a small set of devices to map AI apps and their MCP endpoints, see devices in ~30 minutes, and run mcp-xray.
Tom's Hardware reports Mythos AI reportedly breached 'almost all' NSA classified systems within hours during a red-team test. Learn why teams should deny outbound egress and rotate keys.
Meta says it fixed a bug where Instagram's AI support chatbot could be tricked into changing account recovery emails, a method social posts tie to recent high-profile takeovers.
HiddenLayer's 2026 AI Threat Landscape shows autonomous agents widen the runtime attack surface and account for ~1-in-8 AI breaches. Quick fixes: allowlist, ephemeral tokens, kill switch.