/

CVE-2026-28229

Severity:

CRITICAL RISK

(9.8)

Missing Authorization in Argo Workflows Template Endpoints

CVE-2026-28229

Modified

Vulnerability type

Category:

Missing Authorization

CWE:

CWE-863, CWE-306

Impact:

Sensitive Information Disclosure

Severity:

CRITICAL RISK

(9.8)

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

A missing authorization check in Argo Workflows allows unauthenticated clients to retrieve WorkflowTemplates and ClusterWorkflowTemplates via the template APIs.

When exploited, any party able to send HTTP requests to the Argo Workflows API with a dummy Authorization: Bearer nothing header can access sensitive workflow definitions, including embedded Secret manifests and other confidential configuration data.

Affected context

This vulnerability affects Argo Workflows instances exposed over HTTP/HTTPS that:

  • Run versions prior to 4.0.2 or 3.7.11
  • Expose the WorkflowTemplate and ClusterWorkflowTemplate endpoints
  • Accept or forward arbitrary Authorization headers from clients

It is most relevant in environments where WorkflowTemplates or ClusterWorkflowTemplates contain sensitive configuration, such as embedded Secret manifests, credentials, or internal infrastructure details.

Exploit preconditions

  • Attack surface: Argo Workflows API server (template-related endpoints).
  • Required access: Network reachability to the Argo Workflows API; no valid credentials needed.
  • Trigger point: Sending a request to the WorkflowTemplate or ClusterWorkflowTemplate endpoints using an Authorization: Bearer nothing (or equivalent non-empty) token.
  • Why exploitation is possible: The server-side authorization logic does not correctly enforce access control on these endpoints, treating certain bearer tokens as sufficient without validating identity or permissions.

Why this happens

The core issue is a missing or incomplete authorization check on the WorkflowTemplate and ClusterWorkflowTemplate retrieval endpoints.

Instead of verifying that the caller is authenticated and authorized to read specific templates, the API accepts requests that simply include an Authorization header, allowing arbitrary clients to list or fetch templates.

This maps to CWE-863: Incorrect Authorization, where access control decisions are either skipped or improperly evaluated, resulting in unintended read access to protected resources.

Potential impact

  • Exposure of WorkflowTemplates and ClusterWorkflowTemplates to unauthorized parties
  • Leakage of embedded Secret manifests (e.g., API keys, tokens, passwords) stored in templates
  • Disclosure of internal infrastructure details, such as service endpoints, namespaces, and resource configurations
  • Facilitation of follow-on attacks (e.g., lateral movement, privilege escalation) using leaked credentials
  • Regulatory or compliance impact due to unintended disclosure of sensitive configuration data

General mitigation guidance

Effective mitigation focuses on tightening authorization and minimizing sensitive data exposure in templates:

  • Ensure that all template-related endpoints enforce proper authentication and fine-grained authorization based on the caller's identity and role.
  • Treat WorkflowTemplates and ClusterWorkflowTemplates as sensitive resources and apply least-privilege access controls (e.g., read permissions only to specific service accounts or roles).
  • Avoid embedding raw Secrets or long-lived credentials directly in templates; instead reference external secret management systems (e.g., Kubernetes Secrets, secret managers) with scoped access.
  • Limit network exposure of the Argo Workflows API, placing it behind authenticated gateways or service meshes that enforce strong identity and policy.
  • Adopt secure defaults where anonymous or weakly authenticated access to workflow definitions is disabled.

How teams detect issues like this

Organizations can look for this issue and its symptoms using several approaches:

  • Configuration and code review: Verify that authorization middleware or policy checks are applied consistently to all workflow template endpoints. Look for routes or handlers that bypass access control logic.
  • Static analysis and policy checks: Use tools or custom rules to flag handlers that read or list templates without explicit permission checks or RBAC enforcement.
  • Access log analysis: Inspect API gateway or Argo Workflows logs for template retrieval requests with unusual or obviously invalid Authorization headers (e.g., Bearer nothing) coming from unexpected sources.
  • Security testing: Perform authenticated and unauthenticated API testing to confirm that template endpoints correctly enforce authentication and RBAC policies.

Sources

SourceLink
github.comgithub.com/argoproj/argo-workflows/security/advisories/GHSA-56px-hm34-xqj5
access.redhat.comaccess.redhat.com/errata/RHSA-2026:10184
access.redhat.comaccess.redhat.com/security/cve/CVE-2026-28229
bugzilla.redhat.combugzilla.redhat.com/show_bug.cgi?id=2446549
security.access.redhat.comsecurity.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-28229.json

Frequently asked

questions

Yes. Any attacker who can reach the Argo Workflows API over the network can exploit this issue without valid credentials, provided the vulnerable endpoints are exposed.

No. The vulnerability arises because the presence of an Authorization header is effectively trusted without robust verification, so even a dummy bearer token can be enough to retrieve templates.

Apply strict, centralized authorization checks to all template-related endpoints, ensuring that only authenticated identities with explicit permissions can read or list WorkflowTemplates and ClusterWorkflowTemplates. Combine this with the practice of storing secrets outside of templates.

Adopt a defense-in-depth access control model: enforce mandatory authentication and RBAC on every endpoint, use security testing to validate that no anonymous or weakly authenticated paths exist, minimize secrets in configuration artifacts, and establish code review guidelines that flag any new routes lacking explicit authorization logic.

Find out what's already
exploitable in your codebase.