A concrete proposal for running agent-driven workloads without pretending they are ordinary scripts.
Goal: feedback from Platform on the direction, the boundaries, and what should be validated next.
The concrete names matter less than the responsibility boundaries. The proposed stack puts disposable execution cells at the top, runs them through a self-service platform layer, and lands that platform on a hardened host substrate.
Code: github.com/PSPDFKit-labs/agentic-infrastructure/tree/main/pi-runner-v1
blob.yamlblob deploy/ is wiped/persist, /nix, /var/logA malfunctioning or compromised agent should be able to fail its own task. It should not be able to blow up the host, steal broad credentials, pivot across workloads, or exfiltrate freely.
We need to decide whether this is the right shape to harden.