Summary
Wallos contains an unauthenticated payment-logo search endpoint that passes HTTP proxy environment variables into cURL without disabling proxy use. An attacker may abuse this behavior to cause server-side requests to unintended destinations, including internal services, when the deployment environment exposes attacker-influenced proxy settings.
Affected context
The issue occurs in Wallos deployments using the payment-logo search endpoint at endpoints/payments/search.php before the affected behavior was patched. The endpoint is unauthenticated and uses cURL for outbound requests while allowing HTTP_PROXY or HTTPS_PROXY environment variables to influence proxy routing. Risk depends on the runtime environment, network reachability, and control over those environment variables or related deployment configuration.
Exploit preconditions
- Attack surface: The unauthenticated payment-logo search HTTP endpoint
- Required access: Ability to send requests to the exposed endpoint; additional influence over proxy-related environment configuration may be required depending on deployment
- Trigger point: Outbound cURL processing performed by
endpoints/payments/search.php - Why exploitation is possible: The endpoint passes proxy environment values into cURL without enforcing direct connections or an appropriate allowlist, allowing server-side request routing to be influenced by untrusted or insufficiently controlled configuration
Why this happens
The underlying bug is inconsistent security hardening between near-identical outbound-request endpoints. Although related logo and search functionality disables cURL proxying, the payment search endpoint continues to honor HTTP_PROXY or HTTPS_PROXY through CURLOPT_PROXY. This maps primarily to CWE-918: Server-Side Request Forgery, because application-controlled server requests can be redirected through an attacker-influenced network path. It also reflects confused-deputy behavior and insecure configuration trust, associated with CWE-441 and CWE-1188.
Potential impact
- An attacker may cause requests to pass through an unintended proxy
- The application could reach internal or otherwise restricted network services
- Sensitive response data may be exposed through attacker-controlled infrastructure
- Internal service metadata or credentials could be targeted where network and application controls permit
- Network segmentation assumptions may be weakened
General mitigation guidance
Mitigation generally focuses on making outbound request behavior explicit and reducing trust in deployment-controlled network settings.
Common defensive patterns include:
- Disable proxy inheritance for requests that do not require proxying
- Use a narrowly scoped, centrally managed proxy allowlist when proxy access is necessary
- Avoid deriving security-sensitive request routing directly from mutable environment variables
- Restrict outbound network access from the application using egress policies and least privilege
- Validate destination schemes, hosts, ports, and redirect behavior before making requests
- Keep equivalent endpoints aligned through shared, hardened HTTP-client helpers
- Avoid exposing unnecessary response content from fetched resources
How teams detect issues like this
Organizations can identify similar issues through:
- Static analysis rules for
CURLOPT_PROXY, proxy environment variables, and outbound HTTP clients - Code review checks for unauthenticated endpoints that fetch remote resources
- Searches for inconsistent cURL options across duplicated or near-identical endpoints
- Tests that verify proxy inheritance is disabled or restricted for security-sensitive requests
- Egress monitoring for application requests to private address ranges, unexpected proxies, and unusual destinations
- Policy enforcement requiring outbound destination validation and network-level egress controls
Automated code review and software composition analysis tools may help surface unsafe request-routing patterns, but runtime network monitoring remains important.

