A legitimate user with Lighthouse provider configuration permissions can supply an arbitrary base_url for the openai_compatible provider. The backend then performs outbound HTTP calls that include the API key in the Authorization header. If the supplied URL is attacker‑controlled, the key is sent to an external host, enabling credential leakage and potential further compromise. The flaw is a classic SSRF (CWE‑918) and can affect the confidentiality of credentials and the integrity of internal network resources.

Prowler cloud platform versions prior to 5.33.1 are vulnerable. The issue was fixed in release 5.33.1; all newer releases are not impacted.

The CVSS score of 7.6 classifies the vulnerability as High severity. No EPSS data is available, but the lack of entry in the KEV catalog suggests no public exploits yet. Exploitation requires authenticated API access with provider configuration rights, and the attacker can trigger the request by submitting a crafted POST to /api/v1/lighthouse/providers or /api/v1/lighthouse/providers/{id}/connection. The high score combined with the ability to exfiltrate credentials makes this a significant risk to any organization using the affected Prowler version.