There is no agreed substrate for agentic systems. Every team builds their own orchestration layer, their own permission model, their own audit trail — or skips these things entirely and hopes for the best.
This is the substrate problem. It is not a research problem in the academic sense. It is an engineering problem with research-level difficulty. The decisions made now — in the absence of a standard substrate — will calcify into the infrastructure that future, more capable systems are built on top of.
The kernel analogy
The kernel contract in traditional computing is the agreement between user-space programs and the operating system: what operations are available, how they behave, what guarantees they carry. Stratum is our attempt to define an analogous contract for agentic computing.
We are not claiming to have solved anything. We are claiming that this is the right framing for the problem, and that building in the open is the right approach to developing it.