Arcjet launches a new product called agent runtime security, aimed at helping teams secure AI agents used in production workflows. The offering focuses on visibility into which agents run, the ability to enforce security controls, and maintaining audit trails that show what happened and why.
The company positions the product for both engineering and security/governance teams. Arcjet says it provides observability across agent workflows and supports enforcement of controls before and after “consequential” actions. It also emphasizes governance and compliance evidence, describing the audit capabilities as useful for security reviews and related compliance processes.
Across the two outlets, the main reported details are consistent: this is a product launch by Arcjet for production AI agents, with features spanning discovery of agent activity, runtime control enforcement, and audit trail preservation. Differences in emphasis are limited to wording—one source highlights “audit trails” and workflow-level observability, while the other highlights “agent activity” discovery and evidence for security and compliance review.
