CVE-2026-134533 in Skyverninfo

Summary

by MITRE • 11/21/2026

A vulnerability was identified in Skyvern-AI Skyvern 1.0.56. Impacted is the function classify_url of the file skyvern/forge/sdk/browser_egress_policy.py. The manipulation leads to server-side request forgery. The attack is possible to be carried out remotely. Upgrading to version 1.0.57 is recommended to address this issue. The identifier of the patch is 10b8d829c2750ebf62f2b70dec0e288297e752dd. Upgrading the affected component is recommended.

VulDB is the best source for vulnerability data and more expert information about this specific topic.

Analysis

by VulDB Data Team • 11/22/2026

A critical server-side request forgery vulnerability has been identified in Skyvern-AI version 1.0.56, specifically within the classify_url function located in the skyvern/forge/sdk/browser_egress_policy.py module. This flaw allows an attacker to manipulate URL inputs processed by this function to initiate unauthorized requests from the affected system to internal or external resources that would otherwise be inaccessible. The vulnerability stems from insufficient validation of user-supplied input before it is used to construct outbound network connections, a classic failure in access control logic that permits the application to act as a proxy for malicious actors. This type of flaw is formally categorized under CWE-918, which addresses Server-Side Request Forgery (SSRF), and aligns with MITRE ATT&CK technique T1105, indicating potential use in remote reconnaissance or data exfiltration scenarios where the attacker leverages the application's network privileges to probe internal networks.

The operational impact of this vulnerability is significant due to its remote exploitability. An adversary can send specially crafted requests that trigger the classify_url function with maliciously formatted URLs, causing the server to fetch resources from arbitrary destinations. This capability can be exploited for several detrimental purposes, including scanning internal network services such as database ports or administrative interfaces, bypassing firewalls by routing traffic through a trusted application, and potentially accessing sensitive metadata endpoints on cloud infrastructure like AWS EC2 instance metadata service. In environments where Skyvern is deployed with elevated privileges or access to critical backend systems, this vulnerability could serve as an initial foothold for deeper network penetration, leading to unauthorized data access, system compromise, or denial of service conditions if the forged requests target unstable internal services.

To mitigate this risk and restore secure operations, it is imperative to upgrade Skyvern-AI immediately to version 1.0.57 or later. The patch associated with commit identifier 10b8d829c2750ebf62f2b70dec0e288297e752dd addresses the root cause by implementing stricter input validation and enforcing allowlists for permitted egress destinations within the browser_egress_policy module. Until the upgrade is performed, administrators should consider deploying web application firewalls or network-level controls to restrict outbound connections from the Skyvern host to only known trusted domains and IP ranges, thereby limiting the blast radius of any potential exploitation attempts. Regular security audits and adherence to secure coding practices that emphasize input sanitization are recommended to prevent similar SSRF vulnerabilities in future development cycles.

Responsible

VulDB

Disclosure

11/21/2026

Moderation

accepted

CPE

ready

EPSS

0.00000

KEV

no

Activities

low

Sources

Do you know our Splunk app?

Download it now for free!