Defensible Design for OpenClaw: Securing Autonomous Tool-Invoking Agents

OpenClaw-like agents offer substantial productivity benefits, yet they are insecure by default because they combine untrusted inputs, autonomous action, extensibility, and privileged system access within a single execution loop. We use OpenClaw as an exemplar of a broader class of agents that...

AI and Cybersecurity

Ben Buchanan has written "A National Security Research Agenda for Cybersecurity and Artificial Intelligence." It's really good -- well worth reading...

A New Cybersecurity Research Agenda (In Three Minutes or Less)

Editor’s Note: As the CISO of In-Q-Tel, the CIA-backed strategic investment firm focused on developing technologies for the intelligence community, Dan Geer gets paid to help find the answers to big questions about computer security, national security, privacy and technology. Headlines proliferat...

Some Advice for Howard Schmidt

It’s been little more than 24 hours since President Obama named Howard Schmidt his White House Cybersecurity Coordinator, but it didn’t take nearly that long for just about everyone with any interest in security to line up on one side or the other of the “he has no chance/he’ll save the Internet”...