The validate_url function in crewAI performs a single DNS resolution followed by a blocklist check, but then returns the original URL unchanged. Because the implementation does not examine HTTP redirects or resolve the final destination, an attacker can supply a URL that redirects to an internal address or use DNS rebinding techniques to access internal services and cloud metadata endpoints. This bypass enables the server to make requests to privileged internal resources, potentially exposing sensitive data or facilitating further attacks, and constitutes server‑side request forgery (CWE‑918).

crewAI Inc. crewAI software, versions earlier than 1.15.1 from the 1.15.0 series onward, are affected. The issue is present in all releases that include the Scrape Tools component before the 1.15.1 update.

The vulnerability indicates high severity, with a CVSS score of 8.3. The EPSS score is <1%, and the issue is not listed in the CISA KEV catalog, so the exploitation likelihood is currently unknown. Based on the description, it is inferred that an attacker would trigger the scrape function via an external request—such as an API call—with a crafted URL that redirects to an internal target. If successful, the victim server’s outbound traffic could reach internal services or cloud metadata, compromising confidentiality, integrity, or availability of those resources. The risk assessment therefore remains high, particularly for deployments with exposed public endpoints that utilize the Scrape Tools.