/

CVE-2026-54734

Severity:

CRITICAL RISK

(10.0)

SSRF (CWE-918) in Prebid Server Java Bidder Adapter Requests

CVE-2026-54734

Awaiting Analysis

Vulnerability type

Category:

Server-Side Request Forgery

CWE:

CWE-918

Impact:

Unauthorized server-side network requests and potential access to internal services

Severity:

CRITICAL RISK

(10.0)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

A critical server-side request forgery vulnerability in Prebid Server Java allows certain bidder adapters to build outbound request URLs from attacker-influenced bid parameters without adequately validating the destination.

If an attacker can supply those parameters, the server may be induced to contact internal services, cloud metadata endpoints, or other sensitive destinations reachable from its network.

Affected context

The issue occurs in Prebid Server Java deployments where affected bidder adapters interpolate user-supplied parameters into outbound request URLs without using centralized validation for the resulting domain or path segment. Risk depends on whether untrusted parties can influence bid-request parameters and what internal systems are reachable from the Prebid Server runtime.

Exploit preconditions

  • Attack surface: Bid-request handling and outbound HTTP requests generated by bidder adapters
  • Required access: Ability to submit or influence bid-request parameters processed by an affected adapter
  • Trigger point: Construction and dispatch of an adapter's outbound request
  • Why exploitation is possible: User-controlled URL components can alter the intended destination when domain and path values are not validated against an approved request policy

Why this happens

The underlying bug is a server-side request forgery condition mapped to CWE-918. A bidder adapter incorporates externally controlled values into an outbound URL without applying consistent validation to the resolved host and path.

This insecure pattern allows request construction to cross a trust boundary: data intended to describe a bid request can instead influence where the server connects. Because the request originates from the server, it may have network access and credentials that are unavailable to the original requester.

Potential impact

  • The server may send requests to internal network services
  • Cloud instance metadata or workload identity endpoints could be reached
  • Sensitive responses may be exposed through application behavior or logs
  • Internal administrative interfaces could be probed or abused
  • Network segmentation controls could be bypassed from the server's trusted position
  • Downstream systems could be affected if requests carry credentials or trusted headers

General mitigation guidance

Mitigation generally focuses on enforcing a single, secure outbound-request policy for every adapter.

Common defensive patterns include:

  • Validating the complete resolved destination with centralized HTTP URL utilities
  • Allowing only approved schemes, domains, ports, and path patterns
  • Reducing trust in bid-request fields used to construct URLs
  • Preventing connections to loopback, link-local, private, and otherwise restricted address ranges where business requirements permit
  • Revalidating destinations after hostname resolution and redirect handling
  • Avoiding forwarding ambient credentials or sensitive headers to adapter-controlled destinations
  • Running the service with least-privilege network permissions and isolated egress policies
  • Applying response-size, timeout, and redirect limits to outbound requests

How teams detect issues like this

Detection approaches can include:

  • Static analysis searches for string concatenation, interpolation, or template construction involving user-controlled values and HTTP URLs
  • Code review checks for adapter-specific URL building that bypasses centralized validation utilities
  • Tests that verify rejected schemes, private address ranges, unexpected ports, redirects, and encoded host or path variations
  • Egress monitoring for requests from Prebid Server to internal, metadata, or otherwise unauthorized destinations
  • Network policy alerts for attempted access to restricted address ranges
  • Automated security review tools that identify SSRF-prone request construction and missing destination validation

Sources

SourceLink
github.comgithub.com/prebid/prebid-server-java/commit/a129a685b6fc54441c1a779ce29ea4ec20c2e09f
github.comgithub.com/prebid/prebid-server-java/commit/d0c723ce36aa09712c825ea9625d978f45d29a1c
github.comgithub.com/prebid/prebid-server-java/pull/4522
github.comgithub.com/prebid/prebid-server-java/releases/tag/3.43.0
github.comgithub.com/prebid/prebid-server-java/security/advisories/GHSA-fr2c-g2f8-qchg

Frequently asked

questions

It can be remotely exploitable when an attacker is able to submit bid requests or otherwise control parameters consumed by an affected bidder adapter.

The key requirement is the ability to influence the relevant request parameters. Exploitation does not necessarily require direct access to the Prebid Server host.

Route all adapter-generated requests through centralized destination validation that applies strict allowlists and blocks unauthorized network ranges, schemes, ports, and redirects.

Treat URL components from requests as untrusted, enforce egress restrictions, avoid forwarding credentials, and test every adapter's outbound request behavior against SSRF cases.

Find out what's already
exploitable in your codebase.