The vulnerability is a server‑side request forgery flaw in the Fal.ai media status polling feature of prompts.chat prior to commit 30a8f04. Authenticated users can supply an attacker‑controlled URL in the token parameter, causing the server to issue arbitrary outbound requests. Because Fal.ai URLs are not validated, the request includes the FAL_API_KEY in the Authorization header, exposing the key. This results in credential theft, internal network probing, and abuse of the victim’s Fal.ai account.

The vulnerability is present in prompts.chat deployments running any code prior to commit 30a8f04. Users and administrators should verify the version in use and determine whether the affected code base remains in production.

The flaw has a CVSS score of 7.1, indicating moderate‑to‑high severity, yet the EPSS score is below 1% and it is not listed in the CISA KEV catalog, suggesting a low probability of widespread exploitation. The attack requires authentication to the service, which limits the attack surface, but the potential to exfiltrate sensitive keys and probe internal resources makes it a significant risk if unpatched.