The Orchestrator

The next chapter surveys the tools. Carry these questions into it: what is being developed, what evidence comes back, which standard you are leaning on, and which responsibility that standard leaves with you. [^ch04_oasp_env]: NVIDIA OpenShell policy model — the operator defines which files, networks, tools, processes, and credentials an agent can access; OpenShell turns those limits into a verifiable policy checked before the agent runs and enforced while it works. Sources: NVIDIA technical blog, September 28, 2026, https://developer.nvidia.com/blog/nvidia-open-agent-safety-platform-a-reference-for-continuous-in-silicon-agent-monitoring/; OpenShell repository, https://github.com/NVIDIA/OpenShell. An envelope for resources, not the full machine-readable object this section describes — no authority record, learning policy, or cross-vendor portability. Verified September 29, 2026.

HQ 8 — Authored. I wrote the argument, structure, examples, and prose. AI tools were used for research assistance and light editorial polish; the voice, ideas, and substance are mine.