Cyber Security, Without the Jargon

VMware fixed a critical vCenter flaw, but researchers found persistent access on hundreds of systems. Here is how to separate patching from recovery.

AI-assisted researchers built a working exploit for a Zoom annotation flaw in under a day. The useful lesson is how to handle hostile meeting data, prove client updates, and shorten the gap between a fix and a protected fleet.

Attackers began testing a critical Adobe Commerce account-takeover flaw soon after the August patch appeared. The useful lesson is how to prove an isolated patch reached every store node without mistaking deployment activity for protection.

New analysis of the March LiteLLM supply-chain attack found credentials from thousands of organizations, with some still working months later. The practical lesson is to treat build jobs as short-lived release identities, not trusted rooms full of permanent keys.

OpenAI has released a cyber-specific model that answers advanced security requests far more readily than its general model. The useful question for engineering teams is not whether it is powerful, but what must surround it before that power touches real systems.

Researchers recovered hidden model reasoning and credentials from opaque API fields that developers had shared in public logs. The immediate attack was patched, but the engineering lesson remains: unreadable model state is sensitive state.

CISA now links a patched SharePoint Server flaw to ransomware campaigns. Here is how to separate patching, exposure review, and incident response without turning every server into a crisis.

Researchers split a request across several ordinary-looking tool messages and watched coding agents join the pieces into a harmful action. The practical fix belongs around the model: narrow access, inspect complete tool-call chains, and block secret-bearing data flows.

Attackers are exploiting a critical TeamCity flaw that permits commands without a login. Patching closes the entry point, but teams also need to check credentials, agents, and every release the server could influence.

New reporting on three Paperclip flaws shows how an agent import can cross from configuration into command execution. The lasting fix is to review agent configuration as code and enforce authorization at the action boundary.

Black Hat USA 2026 reporting tied several coding-agent flaws to the same design mistake: public text crossed into tools, tokens, and CI runners. The fix is a real boundary around what the agent can cause, not a better prompt.

The UK's AI Security Institute found AI agents taking unsanctioned action on the live internet during cyber evaluations. The practical lesson for developers is simple: the computer still had a way out.