/

CVE-2026-63506

Severity:

HIGH RISK

(8.8)

Authorization Bypass (CWE-639) in Tina CMS Auth Provider

CVE-2026-63506

Deferred

Vulnerability type

Category:

Authorization Bypass

CWE:

CWE-639

Impact:

Cross-Tenant Unauthorized Access and Content Manipulation

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

Tina CMS contains an authorization flaw that allows a request-controlled client identifier to select which TinaCloud application validates a bearer token. An attacker with any TinaCloud account may use a token from their own application to gain unauthorized access to a victim site's content and media operations.

Affected context

The vulnerability occurs in Tina CMS deployments using the affected authentication or media authorization integrations. It becomes risky when an endpoint accepts a clientID from the request and uses it to choose the TinaCloud application against which the bearer token is validated, rather than consistently using the self-hosted site's configured application.

Exploit preconditions

  • Attack surface: Tina CMS authentication, GraphQL, or media endpoints
  • Required access: A valid account and bearer token for any TinaCloud application
  • Trigger point: An authorization request containing an attacker-selected client identifier
  • Why exploitation is possible: The selected application controls token validation, allowing the attacker's verified status to cross the intended tenant boundary

Why this happens

The underlying bug is an authorization reference flaw mapped to CWE-639, Authorization Bypass Through User-Controlled Key. The authorization logic trusts a clientID supplied by the request when selecting the security context used to validate a bearer token. Because the validated identity is not bound to the self-hosted site's configured TinaCloud application, a token valid in one tenant can be treated as authorized for another tenant.

Potential impact

  • An attacker may list or read victim media
  • An attacker could upload or delete media in the victim site
  • GraphQL read operations may expose victim content
  • GraphQL create or update operations could modify victim content
  • GraphQL delete operations may remove victim content
  • Cross-tenant access could enable content tampering or data loss

General mitigation guidance

Mitigation generally focuses on binding authorization decisions to the

server-side tenant and application configuration rather than request-selected

identifiers.

Common defensive patterns include:

  • Deriving the expected client or application identity from trusted server-side configuration
  • Rejecting client identifiers that do not match the configured tenant context
  • Validating token issuer, audience, tenant, and application claims together
  • Keeping authentication and authorization checks centralized and consistent across GraphQL and media callbacks
  • Applying least privilege to content-management service identities
  • Avoiding request parameters as authoritative selectors for security context

How teams detect issues like this

Organizations can look for this issue through:

  • Static analysis rules that flag request-controlled tenant, client, issuer, or audience selectors in authorization code
  • Code review checks confirming that token validation is bound to server-side configuration
  • Tests attempting cross-tenant token and client-identifier combinations
  • Runtime monitoring for tokens from unexpected applications or issuers
  • Policy enforcement requiring explicit tenant and audience checks before privileged content operations
  • Automated security review tools that trace user-controlled values into authorization decisions

Sources

SourceLink
github.comgithub.com/tinacms/tinacms/commit/0a927a4f8d228dd05ee7ca4be32899bc190e73af
github.comgithub.com/tinacms/tinacms/pull/7168
github.comgithub.com/tinacms/tinacms/releases/tag/@tinacms/auth@1.1.4
github.comgithub.com/tinacms/tinacms/releases/tag/next-tinacms-azure@15.0.1
github.comgithub.com/tinacms/tinacms/security/advisories/GHSA-g74q-6g2f-874x

Frequently asked

questions

Yes. An attacker may exploit affected network-facing Tina CMS endpoints if they possess a valid TinaCloud account and bearer token.

No. The described issue allows an account associated with another TinaCloud application to be accepted across the intended tenant boundary.

Bind token validation to the site's trusted, server-side application configuration and reject request-controlled values that can change the authorization context.

Enforce issuer, audience, tenant, and application consistency checks, test cross-tenant authorization paths, and apply least privilege to content and media operations.

Find out what's already
exploitable in your codebase.