A vulnerable app trusts input (form fields, uploads, API calls).
The attacker crafts input that breaks out of normal handling.
Their code executes—downloading payloads, stealing data, or taking control.
Sudden processes you didn’t start; new services or tasks
Unusual outbound connections right after a form submit/upload
AV/EDR alerts about script hosts, PowerShell, or injected DLLs
Isolate the host; preserve logs and memory.
Block the offending URL/IP at WAF/firewall.
Patch/disable the vulnerable feature; rotate secrets/keys.
Hunt laterally for dropped payloads and persistence.
Validate and sanitize all inputs; treat uploads as untrusted.
Use allow-lists (what’s allowed) rather than block-lists.
Turn on WAF/Runtime protections (e.g., RASP) and strict content security.
Keep apps, libraries, and frameworks updated; avoid risky plugins.
Run services with least privilege; restrict script interpreters.