Approvals Overview

Understand when Noxgild can continue automatically and when an exact approval or human action is required.

Noxgild does not interrupt every ordinary operation with a confirmation prompt. Approval is applied where policy requires it.

Routine work can proceed when account, computer, capability, scope, schema, worker, and policy checks succeed.

Destructive commands and some trust-boundary crossings stop before dispatch and require approval for the exact command. Approval is scoped to that command only.

CAPTCHA, OTP/2FA, passkeys, secure-desktop prompts, and similar protected challenges are not Noxgild approvals. They remain human actions.

Continue with How approvals work for the full policy model.