Agent governance stops at the company boundary — exactly where agent commerce begins.
Every standard control — identity, restriction, logging — presumes the agent runs inside a perimeter its deployer controls, so cross-company agent-to-agent action is governed only by emergent mutual authentication, shared logging standards, and platform blocks rather than by law, insurance, or complete audit trails.
The same conclusion keeps arriving from across the workspace's research — 1 topics independently instantiate this theme. Filter the evidence by where it came from:
Identity ambiguity, cascading actions, and fractured audit trails appear precisely at the cross-organizational seam where liability disperses and insurance excludes the loss.
The first binding enforcement against a consumer agent came from a counterparty platform applying its own boundary rules — Apple and Amazon, not regulators — a consumer-scale instance of unpoliced agent crossings.