The vulnerability in modelscope agentscope (up to version 1.0.18) involves improper handling of URL arguments in certain functions related to image and audio processing. This flaw enables an attacker to craft malicious URLs that cause the server to make unintended requests, leading to server-side request forgery. The affected functions are _parse_url, prepare_image, and openai_audio_to_text within the Cloud Metadata Endpoint component. The vulnerability can be exploited remotely without authentication and requires no user interaction. The vendor has not provided any advisory or patch, and the exploit is publicly available.

Successful exploitation of this SSRF vulnerability could allow an attacker to make arbitrary requests from the vulnerable server, potentially accessing internal resources or services not otherwise reachable. The CVSS score of 6.9 reflects a medium severity impact with low complexity and no privileges required. There is no indication of known exploits in the wild at this time. The lack of vendor response or patch means the vulnerability remains unmitigated in affected versions.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Since the vendor has not responded or provided a fix, users should consider implementing network-level controls to restrict outbound requests from the affected component or apply input validation and sanitization to URL parameters as a temporary mitigation. Monitor for any updates from the vendor regarding official patches or fixes.