/

CVE-2026-88009

Severity:

HIGH RISK

(8.2)

HTTP Request Smuggling (CWE-444) in Traefik HTTP Proxy

CVE-2026-88009

Analyzed

Vulnerability type

Category:

HTTP Request Smuggling

CWE:

CWE-444, CWE-1286

Impact:

Routing and Authorization Bypass

Severity:

HIGH RISK

(8.2)

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 request-parsing discrepancy in Traefik can cause HTTP/1 request targets with a rootless form to be normalized differently for routing and backend forwarding. An attacker may use this mismatch to bypass cross-vhost or path-based authorization controls and evade access logging when the backend interprets the forwarded opaque target as a path.

Affected context

The issue affects Traefik HTTP reverse-proxy deployments handling specially formed HTTP/1 request targets. It becomes risky when Traefik evaluates routing, sanitization, authentication forwarding, encoded-character policies, or logging using a normalized path while the backend receives the original opaque target and interprets it differently.

Exploit preconditions

  • Attack surface: HTTP/1 request handling at a Traefik reverse-proxy entry point
  • Required access: Ability to send requests to an exposed Traefik router
  • Trigger point: Processing of a rootless HTTP/1 request target stored by Go in URL.Opaque
  • Why exploitation is possible: Traefik performs security and routing decisions using a normalized / path but forwards the opaque target verbatim, creating inconsistent interpretations between the proxy and backend

Why this happens

The vulnerability results from inconsistent parsing and normalization of an HTTP request target across proxy processing stages. Traefik's routing, path sanitization, authentication forwarding, encoded-character handling, and access logging evaluate a normalized path, while backend forwarding preserves URL.Opaque. When the backend interprets that opaque value as a path, the proxy and backend disagree about the requested resource.

This behavior maps primarily to CWE-444: Inconsistent Interpretation of HTTP Requests, with CWE-1286: Improper Validation of Syntactic Correctness of Input also relevant because an unusual request-target form is accepted without ensuring consistent semantic validation across all processing stages.

Potential impact

  • An attacker may bypass cross-vhost routing restrictions
  • Path-scoped authorization checks could be bypassed
  • Protected backend resources may become reachable through inconsistent path interpretation
  • Access logs may omit or misrepresent the effective backend path
  • Monitoring and incident investigation could be impaired
  • Backend behavior could be influenced by request-target parsing differences

General mitigation guidance

Mitigation generally focuses on enforcing one consistent interpretation of HTTP request targets throughout proxy processing.

Common defensive patterns include:

  • Rejecting malformed, ambiguous, or unsupported request-target forms at the earliest boundary
  • Applying identical canonicalization and validation rules to routing, authorization, forwarding, and logging
  • Avoiding decisions based on a normalized placeholder when a separate raw target will be forwarded
  • Ensuring backend services receive only validated, canonical request paths
  • Treating proxy and backend parser disagreement as a security policy violation
  • Restricting backend privileges and isolating sensitive routes so parser discrepancies have limited impact
  • Testing reverse-proxy behavior with differential request-target and normalization cases

How teams detect issues like this

Detection can combine parser-focused testing, source review, and runtime monitoring.

  • Use static analysis and code review to identify separate handling of normalized path fields and raw or opaque request-target fields.
  • Review whether routing, authorization, forwarding, and logging consume different representations of the same request.
  • Add regression tests for rootless HTTP/1 targets, unusual request-target forms, encoded separators, and proxy/backend interpretation differences.
  • Enforce gateway policies that reject ambiguous request syntax before routing decisions are made.
  • Compare proxy access logs, backend logs, and observed request paths for mismatches.
  • Neutral automated review tools may help identify inconsistent canonicalization and HTTP parser-boundary patterns.

Sources

SourceLink
github.comgithub.com/traefik/traefik/commit/58d1e9ca204526823211e30fd4634101c59d58e9
github.comgithub.com/traefik/traefik/pull/13796
github.comgithub.com/traefik/traefik/releases/tag/v2.11.57
github.comgithub.com/traefik/traefik/releases/tag/v3.7.13
github.comgithub.com/traefik/traefik/security/advisories/GHSA-f52w-8j3h-j724

Frequently asked

questions

It may be exploitable remotely when an attacker can send HTTP/1 requests to a Traefik entry point that forwards traffic to a backend interpreting opaque request targets as paths.

Not necessarily. The primary requirement is access to an affected HTTP route; the resulting impact depends on routing rules, authentication controls, and backend behavior.

Ensure ambiguous request-target forms are rejected or canonically interpreted once, then use the same validated representation for routing, authorization, forwarding, and logging.

Maintain consistent parser behavior across proxy and backend components, add differential HTTP request tests, and monitor for discrepancies between proxy records and backend-observed paths.

Find out what's already
exploitable in your codebase.