Kevin Lin describes starting OpenClaw Enterprise at OpenAI, donating the project to the OpenClaw Foundation and working with Red Hat and NVIDIA. He presents it as an open, vendor-neutral platform for managing persistent agents in enterprise environments.
Lin discusses separating trusted configuration and policy enforcement from agent workloads, with the control planeAn AI control plane observes AI activity and applies centralized policy without owning the application's primary task logic. running across an isolation boundary. The conversation covers role-based access, audit logs, sandbox integrations and a common contract for tools such as gateways and secret brokers.
The speakers describe a pre-1.0 project with pilots and planned work on identity, authentication and integration contracts. Lin suggests starting with limited or read-only accessAn agent permission boundary defines and enforces which resources and actions an automated agent is allowed to use. before considering more capable coding agents; these are design intentions and early experiences, not a guarantee of security.
Watch on YouTube
