Clarify Hyperkernel application platform positioning

[hnordt]

What changed

• Reframed the package description and keywords around Hyperkernel as an application platform for modular applications.
• Revised the README opening to establish the product hierarchy: Hyperkernel is the shared platform, applications are built and run on it, and those applications work together as a software platform tailored to an organization’s work.
• Removed the remaining foundation and shared-kernel-primitives framing from the README’s examples so the hierarchy stays consistent throughout the document.
• Integrated that hierarchy with the existing kernel and fast-moving application trust zones.
• Updated only the opening product definition in the AGENTS.md Mission section so the engineering guidance uses the same model.
• Preserved the README Status section’s distinction between the target architecture and the current implementation.

Why

The previous “primitives for building software platforms” wording was ambiguous because it could make Hyperkernel sound like a toolkit used to construct a separate platform. It also placed the implementation mechanism, the kernel and its primitives, ahead of the product itself.

The resulting developer-facing model is explicit: Hyperkernel is the self-hostable application platform. Developers and organizations build and run modular applications on Hyperkernel. Those applications share contracts for identity, permissions, state, events, and audit history, and collectively form a software platform tailored to the work they support.

The event-sourcing keyword was removed because the production event store and replay architecture remain target architecture rather than implemented functionality.

Validation

• git diff --check
• npm run lint