Summary
A server-side request forgery vulnerability in WuzhiCMS allows a remote attacker to influence the destination requested by the Remote Image Fetch feature.
When the affected source[] input is processed by ckditor::saveRemote, the server may be induced to send requests to attacker-selected or internal resources.
Affected context
The issue occurs in WuzhiCMS deployments using the Remote Image Fetch functionality and the ckditor::saveRemote handler in coreframe/app/attachment/index.php. Risk depends on whether the endpoint is remotely reachable, whether attacker-controlled source[] values are accepted, and what network resources and credentials are accessible from the application server.
Exploit preconditions
- Attack surface: The remotely reachable Remote Image Fetch endpoint
- Required access: Access sufficient to submit requests to the affected handler; authentication requirements may depend on deployment configuration
- Trigger point: Processing of attacker-controlled values supplied through the
source[]argument - Why exploitation is possible: The server uses supplied destination data when making outbound requests without adequately restricting reachable schemes, hosts, ports, or network locations
Why this happens
The vulnerability results from insufficient separation between untrusted request data and server-controlled network destinations. The saveRemote functionality appears to trust attacker-influenced source values when fetching remote images, allowing the application server to act as a proxy to destinations that the requester may not be able to access directly. This behavior maps to CWE-918: Server-Side Request Forgery.
Potential impact
- Requests to internal or otherwise restricted services could be made from the application server
- Cloud or service metadata could be exposed where reachable and insufficiently protected
- Internal administrative interfaces could be probed or accessed
- Sensitive response data could potentially be returned to or processed by the application
- Network segmentation assumptions could be weakened
General mitigation guidance
Mitigation generally focuses on making outbound fetching narrowly scoped and treating all remote image locations as untrusted.
Common defensive patterns include:
- Allowing requests only to explicitly approved schemes, hosts, ports, and paths
- Resolving hostnames and rejecting loopback, link-local, private, reserved, and otherwise restricted address ranges
- Revalidating destinations after DNS resolution and across redirects
- Applying outbound network policies, egress filtering, and dedicated fetcher isolation
- Avoiding unnecessary transmission of credentials or sensitive headers during outbound requests
- Limiting response size, request duration, and redirect depth
- Running the image-fetching process with minimal network and filesystem privileges
How teams detect issues like this
Detection approaches can include:
- Static analysis for request handlers that pass user-controlled URLs into HTTP clients or URL-opening APIs
- Code review checks for unrestricted redirects, hostname validation gaps, and private-address filtering failures
- Web application logs monitoring unusual outbound destinations, internal IP ranges, metadata endpoints, and unexpected protocols
- Egress firewall and network policy alerts for application containers contacting restricted network segments
- Policy enforcement requiring approved URL parsers, host allowlists, and bounded outbound request behavior
- Neutral automated security review tools that identify user-controlled network requests and SSRF-prone data flows

