ROE v1 ships with four standing templates. They are designed like reference pages: scope, doctrinal basis, attack scenarios, and policy text in one place.
Escalates or blocks money movement, refunds, and payment changes unless an operator has explicitly authorized the action path.
Positive Identification (PID), proportionality, and standing vs. special ROE.
Restricts outbound automations such as webhooks, mass messaging, and third-party workflow execution unless the workflow is already sanctioned.
Restricted engagement zones and escalation of force, translated into outbound automation controls.
Blocks destructive filesystem actions by default and escalates risky write/delete operations based on blast radius.
Collateral damage estimation and proportionality for local system changes.
Lets Hermes/Discord agents inspect thread context, escalates verified single-thread cleanup, and denies broad or parent destructive cleanup.
Positive identification and restricted engagement zones translated into Discord thread cleanup authority boundaries.
Allows agents to read, draft, and preview communications while requiring human approval before email, chat, ticket, recruiter, social, or webhook messages leave the system.
Positive identification, proportionality, collateral damage estimation, and escalation of force for reputation-bearing external messages.