SECURITY Signal 462
A new security baseline for enterprise agentic adoption
↗
Agent Baseline is an open blueprint by Docker, Snyk, and Keycard defining 35 controls across six security outcomes for deploying enterprise AI agents without unchecked authority.
AI agents can be reprogrammed at runtime through natural-language instructions, operate under delegated credentials, and spawn sub-agents faster than humans can review, turning familiar security controls into a new systems problem. The baseline gives security teams a minimum set of outcomes to govern these agents rather than relying on model-level safeguards alone.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
Agent Baseline v1.0 draft defines 35 controls across six outcomes: Discover, Constrain, Authorize, Observe, Validate, and Respond.
The blueprint treats agents as runtime-programmable actors that can be redirected via natural language, delegate credentials, and spawn sub-agents faster than human review permits.
Docker, Snyk, and Keycard created the baseline and launched it at Black Hat 2026.
THE CLUSTER