Current multi-agent authorization architectures prioritize system integrity over human outcomes. While recent models successfully prevent privilege escalation, they externalize the cost of failure onto the human operator. Re-centering authorization around human impact exposes four structural vulnerabilities in current paradigms. First, calling an agent's understanding of user intent verified is misleading. Intent has no formal definition, so no check against it can be definitive. Systems should instead report an openly probabilistic estimate and scale their caution with the risk of the action. Second, we argue that dynamic trust scores do not survive multi-agent handoffs; transmitting algorithmic judgment across a company boundary without underlying evidence risks violating Contextual Integrity. Third, treating a safely halted workflow as a successful refusal is a category error if it strands the user; compensating actions must be bound directly to the capability token, not only to the planning layer. Finally, we hypothesize that frictionless automated recovery introduces a moral hazard, disincentivizing developers from building robust initial intent checks. We conclude by proposing an architectural synthesis to bridge the gap between deterministic security and human outcomes, alongside the fundamental theoretical limits that remain unresolved.
翻译:暂无翻译