Stored injection is an attack pattern where malicious input is saved by one part of a system and executed later by another. The payload may be written into a database, configuration field, log, profile, or queue, then reused without proper validation or escaping. Because the dangerous content survives the first step, the attack can be persistent and may trigger long after the original request.
This matters in cyber security because the second process often runs with different privileges or in a different context. For example, a router or web admin tool might store a value such as a device name, parental-control rule, or VPN parameter and later pass it into a shell command or parser. Defenders reduce the risk by treating all stored input as untrusted, validating data at ingestion, encoding it when rendering or executing, and avoiding command shells for routine tasks. Stored injection is common in web apps, management consoles, and automated workflows.