PraisonAI contains a logic flaw in its URL checking routine that allows attackers to craft requests that bypass safeguards, leading to SSRF. This flaw permits the application to resolve arbitrary URLs sent by a user, potentially causing the system to send requests to internal network addresses or other unintended destinations. The flaw is categorized as CWE-918 and can lead to the compromise of confidentiality and integrity of internal resources accessible through the application.

The vulnerability affects PraisonAI by MervinPraison, specifically all releases prior to version 1.6.32. Any deployment running a version older than 1.6.32 is considered vulnerable and should be evaluated for an upgrade.

The CVSS score of 7.7 classifies this as high severity; the EPSS score is not available, so the current exploitation probability is unknown. It is not listed in the CISA KEV catalog. Attackers can exploit this remotely by submitting crafted HTTP requests to the application’s URL handling endpoint, surreptitiously making the application perform internal network requests.