Executive brief
Zeep, a Python library used for communicating with SOAP-based web services, failed to enforce security settings designed to block external web requests during document parsing. An attacker who can influence the web service definitions (WSDL or XSD files) loaded by the application can force the server to make unauthorized requests to internal or sensitive web addresses. This could lead to the exposure of internal network information or sensitive cloud metadata.
Technical details
A Server-Side Request Forgery (SSRF) vulnerability exists in Zeep due to a regression introduced in version 4.0.0 where the 'Settings.forbid_external' configuration was defined but not implemented in the parsing logic. When Zeep parses WSDL or XSD documents, it transitively follows 'xsd:import', 'xsd:include', 'wsdl:import', and lxml entity/DTD references. An attacker who can provide or influence a WSDL/XSD document can point these references to internal-only services or cloud metadata endpoints (e.g., 169.254.169.254). The vulnerability is exploited when the application loads the malicious document, causing the server to issue outbound requests to attacker-chosen destinations. The issue is fixed in version 4.3.3 by correctly wiring the 'forbid_external' setting to block these transitive remote loads.
Affected products
- mvantellingen zeep >= 4.0.0, < 4.3.3
Timeline
- 2026-06-18: patched: Fix committed and version 4.3.3 released.
- 2026-06-19: advisory: GitHub Security Advisory published.
- 2026-07-08: disclosed: CVE-2026-58501 published to NVD.