Skip to main content
Vulnerability Database/CVE-2026-59358

CVE-2026-59358: Cloud Foundry UAA Auth Bypass Vulnerability

CVE-2026-59358 is an authentication bypass flaw in Cloud Foundry UAA that lets attackers escalate user tokens to privileged client credentials. This article covers technical details, affected versions, and mitigation.

Published:

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

CVE-2026-59358 Overview

Cloud Foundry User Account and Authentication (UAA) service contains an improper authentication flaw [CWE-287] in its OAuth token endpoint. The client_credentials grant handler does not verify that the Bearer credential presented for client authentication is an actual client secret. UAA accepts any valid access token whose client_id matches the request, including user tokens obtained through public authorization_code + PKCE flows. An authenticated attacker can replay their own user access token in place of a client secret and receive a new token carrying the client's full authorities, such as clients.write. This enables arbitrary OAuth client creation without ever possessing the real client secret.

Critical Impact

Authenticated attackers can escalate from a uaa.user-scoped user token to a fully-privileged client_credentials token, enabling creation of attacker-controlled OAuth clients with arbitrary authorities.

Affected Products

  • Cloud Foundry UAA (User Account and Authentication service)
  • OAuth clients configured for both public authorization_code and client_credentials grant types on the same client_id
  • Cloud Foundry deployments relying on UAA for OAuth client administration

Discovery Timeline

  • 2026-10-06 - CVE-2026-59358 published to NVD
  • 2026-10-06 - Last updated in NVD database

Technical Details for CVE-2026-59358

Vulnerability Analysis

UAA's token endpoint conflates two distinct authentication contexts. When processing a client_credentials grant request, the endpoint must verify a client secret or a configured client authentication method. Instead, the handler accepts any Bearer-presented access token whose client_id matches the request body.

A user token obtained through the standard authorization_code + PKCE flow carries a user_id, is scoped only to uaa.user, and records client_auth_method=none. These attributes mark it as a user credential, not a client credential. The token endpoint ignores these attributes and treats the token as sufficient proof of client identity.

UAA then issues a fresh client-only token populated with the full authorities configured on the OAuth client. If the client holds clients.write, the attacker gains the ability to create arbitrary new OAuth clients with attacker-chosen authorities. The original user token cannot call POST /oauth/clients directly (that endpoint correctly returns 403), so the escalation path depends entirely on the token swap at the /oauth/token endpoint.

Root Cause

The client_credentials code path performs no type check on the credential supplied for client authentication. It verifies that the Bearer token is cryptographically valid and that its client_id matches the request, but it does not verify that the token represents a client-authentication artifact (secret, JWT bearer assertion, or mTLS binding). Any access token issued by UAA for the same client satisfies the check.

Attack Vector

Exploitation requires network access to the UAA token endpoint and a valid user access token for a target OAuth client. The client must support both a public user-facing authorization flow and the client_credentials grant type on the same client_id, which is a non-default configuration.

The attacker completes a normal authorization_code + PKCE flow against the target client, obtaining a user-scoped access token. They then submit a POST /oauth/token request with grant_type=client_credentials and client_id matching the client, presenting the user token as Authorization: Bearer <user_token> instead of HTTP Basic client credentials. UAA responds with a new token carrying the client's authorities. The attacker uses that token to call POST /oauth/clients and provision new clients with authorities of their choosing.

Detect Identity Attacks the Moment Authentication Fails

Singularity™ Identity surfaces credential misuse, lateral movement, and privilege escalation in real time, so you can contain the breach before it spreads.

Stop Identity Attacks

Detection Methods for CVE-2026-59358

Indicators of Compromise

  • Successful POST /oauth/token requests with grant_type=client_credentials where the Authorization header carries a Bearer token rather than Basic client credentials
  • Issuance of client_credentials tokens shortly after an authorization_code exchange for the same client_id
  • Unexpected POST /oauth/clients calls creating new OAuth clients with elevated authorities such as clients.write, clients.admin, or scim.write
  • UAA audit events showing client_auth_method=none associated with token exchanges that produced client-only tokens

Detection Strategies

  • Correlate UAA token-endpoint logs for the same client_id where a user token issuance is followed by a client_credentials token issuance from the same source IP within a short window
  • Alert on any client_credentials grant response that was authenticated using a Bearer token containing a user_id claim
  • Baseline OAuth client inventory and alert on new client registrations, especially those requesting administrative authorities

Monitoring Recommendations

  • Forward UAA audit logs (/uaa/logs) and token endpoint access logs to a centralized SIEM for correlation
  • Monitor for growth in the OAuth client table and review clients.write usage weekly
  • Track per-client grant-type usage and flag clients that exercise both authorization_code and client_credentials grants from end-user sources

How to Mitigate CVE-2026-59358

Immediate Actions Required

  • Inventory all UAA OAuth clients and identify any configured with both authorization_code and client_credentials grant types on the same client_id
  • Split such clients into separate client_ids: one for user-facing authorization flows and one for client_credentials
  • Rotate client secrets and revoke outstanding tokens for any affected client
  • Audit the OAuth client table for unauthorized client registrations created since the vulnerability was disclosed

Patch Information

Refer to the Cloud Foundry Blog CVE-2026-59358 Analysis for the fixed UAA release and upgrade guidance. Operators should upgrade UAA to the patched version and redeploy the Cloud Foundry environment.

Workarounds

  • Remove the client_credentials grant type from any OAuth client that also exposes a public user-facing authorization flow
  • Restrict the authorities granted to OAuth clients so that no client that participates in public authorization flows holds clients.write or similar administrative scopes
  • Enforce client authentication methods other than none (for example, client_secret_basic, private_key_jwt, or mTLS) for any client that uses client_credentials
  • Place UAA token endpoints behind network controls that log and inspect Authorization headers for anomalous Bearer-for-client-auth patterns

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.