An outbound network policy defines which external destinations a process, application, or device is allowed to contact. Rules may restrict connections by destination address, domain, port, protocol, or other criteria. Unlike controls on incoming traffic, these rules govern connections initiated from inside an environment.
In cybersecurity, limiting outbound access can reduce data exfiltration, block communication with attacker-controlled infrastructure, and prevent software from reaching unapproved services. For AI coding agents, a restrictive policy can help contain a compromised or misbehaving process by limiting where it can send data or fetch code. Defenders should verify the policy is enforced at the network or runtime layer, and allow only destinations needed for the task.