Report: Full Disclosure mailing list, 2026-10-31,
https://seclists.org/fulldisclosure/2026/Nov/6
Disposition: VALID-HARDENING, WW-5855 (THREAT_MODEL.md sections 9, 13)
TLP:CLEAR

In short

The comparison is correct but not constant-time, and the project is
changing it. It is not a vulnerability: the token is a 165-bit random value
held server-side per session, the interceptor is documented as protection
against duplicate submissions, and CSRF defence is the application's
responsibility in the threat model.

What the report describes

TokenHelper.validToken compares the submitted token with the session token
using String.equals. The post argues that response timing reveals how many
leading characters match and estimates that about 1,150 requests could
recover a token. No measurements are given.

What the framework does with that input

- The token is generated from a SecureRandom 165-bit integer rendered in
base 36 (TokenHelper.generateGUID) and stored in the session.
- validToken compares the submitted value with the stored one using
String.equals; on a mismatch the session token is kept, on a match it is
removed (TokenHelper.validToken).
- The token interceptor is documented as ensuring that one request per
token is processed, to guard against back buttons and double clicks (token
interceptor page).

Why this is not a vulnerability

The threat model (section 9) places CSRF token management with the
application; the framework token is a double-submit guard. The project has
no evidence that a per-character timing difference in a JVM string
comparison is measurable across a network against a per-session random
token, and the report provides none. The comparison is nevertheless being
made constant-time, as a hardening change.

What the project is doing

WW-5855 replaces the comparison with a constant-time one.

What users should do

Nothing is required. Applications that need CSRF protection should use a
dedicated CSRF mechanism, as the security guide describes.

References

- https://issues.apache.org/jira/browse/WW-5855
- https://struts.apache.org/core-developers/token-interceptor.html
- https://struts.apache.org/security/
- https://github.com/apache/struts/blob/main/THREAT_MODEL.md (sections 9,
13)
- https://github.com/apache/struts/blob/main/SECURITY.md

New evidence goes to [email protected]; questions about this
notice go to this list.

on behalf of the Apache Struts PMC
Ɓukasz

Reply via email to