deciding-who-decides

To determine who decides, you evaluate the action's reversibility, evidence quality, and consequence. This process matches these factors against a documented decision-rights line to place authority where competence and evidence sit.

An escalation trigger occurs when an action is irreversible, trust-bearing, or thinly evidenced. These criteria force explicit escalation to a named human gate to prevent bypassing oversight on high-consequence changes.

You need a decision-rights line when an action may be irreversible or high-trust. Defining this line ensures that edge agents bypass mandatory human gates only when evidence is sufficient and consequences are reversible.

Evaluating edge versus gate placement requires four inputs: the concrete action, its reversibility, observed evidence, and its consequence. These inputs determine if an edge agent can proceed or must escalate.

The best way to prevent irreversible edge changes is to document a decision-rights line with mandatory human gates. This enforces explicit escalation based on reversibility and consequence before action execution.