/

CVE-2026-39944

Severity:

HIGH RISK

(8.8)

CBC Token Tampering (CWE-327) in Ceph RGW STS Endpoint

CVE-2026-39944

Awaiting Analysis

Vulnerability type

Category:

Improper Cryptographic Implementation

CWE:

CWE-327

Impact:

Privilege Escalation to RGW Administrative Access

Severity:

HIGH RISK

(8.8)

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

Ceph RADOS Gateway uses unauthenticated AES-128-CBC encryption to protect STS session tokens, allowing a token holder to alter authorization-related fields without detection. An attacker with any valid STS token may forge administrative privileges and access RGW operations that should require elevated permissions.

Affected context

The vulnerability occurs in Ceph deployments where the RADOS Gateway exposes the S3-compatible endpoint with STS enabled and uses session tokens protected by unauthenticated CBC encryption. It becomes risky when an attacker can obtain any valid STS token, including one without administrative privileges, because token integrity is not verified before authorization decisions are made.

Exploit preconditions

  • Attack surface: Remote access to the RGW S3 endpoint
  • Required access: A single valid STS session token; elevated privileges are not required
  • Trigger point: Token processing during authenticated RGW requests
  • Why exploitation is possible: AES-CBC encryption does not authenticate ciphertext, allowing controlled modification of token fields without requiring an encryption oracle or network observation

Why this happens

The underlying flaw is the use of encryption without an integrity mechanism for security-sensitive authorization data. AES-CBC can provide confidentiality but does not detect deliberate ciphertext changes; predictable changes to corresponding plaintext fields may therefore go unnoticed. This maps to CWE-327: Use of a Broken or Risky Cryptographic Algorithm, because the cryptographic construction is unsuitable for protecting authorization claims without authenticated encryption or an equivalent integrity check.

Potential impact

  • An attacker could escalate a regular STS session to RGW administrative access.
  • Forged authorization fields could bypass capability checks.
  • An attacker may read, modify, or delete accessible object-storage data.
  • Administrative control could enable broader storage-service disruption or data exposure.
  • Compromised credentials could support persistence or lateral abuse of connected systems.

General mitigation guidance

Mitigation generally focuses on ensuring that authorization-bearing tokens are both confidential and integrity-protected.

Common defensive patterns include:

  • Using authenticated encryption or a cryptographically strong, independently verified token-integrity mechanism.
  • Rejecting tokens whenever authentication tags, signatures, claims, or structural validation fail.
  • Avoiding client-controlled authorization fields and deriving privileges from trusted server-side state.
  • Applying least privilege to STS identities, service accounts, and administrative interfaces.
  • Limiting token lifetime, scope, and audience to reduce the impact of token compromise.
  • Rotating or invalidating potentially exposed sessions through established credential-management processes.

How teams detect issues like this

Organizations can identify similar issues through:

  • Static analysis and cryptographic API reviews that flag CBC encryption without authentication.
  • Code review checks for tokens containing privilege or identity claims without signatures or authentication tags.
  • Tests that mutate encrypted token bytes and verify reliable rejection.
  • Policy enforcement requiring authenticated encryption or approved signing libraries for security-sensitive tokens.
  • Monitoring for unusual administrative actions performed by low-privilege STS identities and sudden changes in token-related access patterns.

Sources

SourceLink
github.comgithub.com/ceph/ceph/releases/tag/v19.2.6
github.comgithub.com/ceph/ceph/releases/tag/v20.2.4
github.comgithub.com/ceph/ceph/security/advisories/GHSA-j73r-qrgx-jvq2

Frequently asked

questions

Yes. Exploitation can be performed through the RGW S3 endpoint when STS is enabled and the attacker possesses a valid STS token.

No. A normal valid STS token is sufficient, and the attack does not require an encryption oracle or observation of network traffic.

Protect authorization-bearing tokens with authenticated encryption or a robust integrity mechanism, and reject any token whose integrity or claims validation fails.

Treat encrypted tokens as untrusted input, prohibit unauthenticated encryption for privilege data, apply least privilege, and test that tampered tokens cannot alter authorization outcomes.

Find out what's already
exploitable in your codebase.