/

CVE-2026-23944

Severity:

CRITICAL RISK

(9.8)

Missing Authentication in Arcane Docker Management Interface

CVE-2026-23944

Analyzed

Vulnerability type

Category:

Missing Authentication

CWE:

CWE-306

Impact:

Unauthenticated Access to Remote Environment Resources

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

Arcane's environment proxy middleware forwarded certain /api/environments/{id}/... requests to remote environment agents without enforcing authentication first.

When the environment ID pointed to a remote environment, the proxy attached a valid manager-held agent token even for unauthenticated callers, allowing direct access to remote containers, logs, and other agent operations.

Affected context

This vulnerability affects Arcane deployments that manage one or more remote environments via agents and expose the Arcane HTTP API.

Specifically, it occurs when:

  • Remote environments are reachable through the manager's environment proxy middleware
  • The API path pattern /api/environments/{id}/... is used to access remote resources
  • Authentication is expected to be enforced globally, but proxy logic evaluates environment routing before auth checks

The issue does not depend on any particular container runtime configuration on the remote hosts, only on the presence of Arcane's agents and manager.

Exploit preconditions

  • Attack surface: Arcane HTTP(S) API endpoint exposed on the network (LAN or Internet).
  • Required access: Network connectivity to the Arcane manager; no valid user credentials required.
  • Trigger point: Any HTTP request to paths matching /api/environments/{remote-id}/... targeting a non-local (remote) environment.
  • Why exploitation is possible: The environment proxy middleware decides to forward the request to a remote agent before authentication is enforced and automatically appends the manager-held agent token, effectively authenticating the attacker's request to the agent.

Why this happens

The core issue is a missing authentication check on a security-sensitive endpoint, mapped to CWE-306: Missing Authentication for Critical Function.

The request routing logic in the environment proxy middleware performs the following insecure pattern:

  • It inspects the environment ID to determine whether the request targets a local or remote environment.
  • For remote environments, it immediately proxies the request to the corresponding agent and injects the manager's agent token.
  • Authentication and authorization checks occur after this proxy decision, meaning unauthenticated requests are treated as trusted when reaching the remote agent.

This inversion of responsibility—routing and credential injection preceding access control—creates an implicit trust path that bypasses the intended authentication layer.

Potential impact

  • Unauthenticated listing of containers and services in remote environments.
  • Reading sensitive data from container logs or other agent-backed endpoints.
  • Performing unauthorized management operations on remote environments (e.g., starting, stopping, or removing containers, images, networks, or volumes) where exposed via the agent.
  • Service disruption by stopping or reconfiguring workloads in remote environments.
  • Potential pivoting into broader infrastructure via compromised containers or services.

General mitigation guidance

Effective mitigation focuses on enforcing authentication and authorization before any proxying or credential injection occurs:

  • Ensure that all environment-related API routes, especially /api/environments/{id}/..., are protected by mandatory authentication middleware that runs before routing or proxy logic.
  • Design proxy middleware to assume requests are untrusted by default and require upstream security context (authenticated user identity and roles) before attaching internal tokens or forwarding to agents.
  • Apply least-privilege to agent tokens and internal credentials used for remote environments, limiting the scope of actions they can perform and segregating environments where feasible.
  • Introduce defense-in-depth access controls at the remote agent side (e.g., verifying the caller identity or enforcing additional checks beyond a single bearer token).
  • Restrict external exposure of the Arcane API with network-layer controls such as VPNs, IP allowlists, or API gateways that enforce authentication.

How teams detect issues like this

Detection can rely on a combination of code-level and operational signals:

  • Code review and static analysis:
  • Search for routes under /api/environments/{id}/... that perform proxying or request forwarding.
  • Flag middleware that injects credentials (tokens, API keys) prior to verifying authentication and authorization.
  • Enforce architectural policies where authentication middleware must be at the top of the handler chain for all externally reachable routes.
  • Configuration and policy checks:
  • Use API gateway or service mesh policies to require authentication on all environment management endpoints.
  • Validate that anonymous or guest roles have no access to management APIs.
  • Runtime monitoring:
  • Monitor Arcane API access logs for unauthenticated requests to /api/environments/{id}/... paths and correlate with actions taken by remote agents.
  • Alert on anomalous environment operations originating from IP ranges or user agents that typically do not manage infrastructure.

Automated security tooling (SAST, API security scanners, or gateway rule validators) can help identify routes where authentication is skipped or applied conditionally.

Sources

SourceLink
github.comgithub.com/getarcaneapp/arcane/commit/2008e1b93b25d0c4c3fff3af07843766231614eb
github.comgithub.com/getarcaneapp/arcane/pull/1532
github.comgithub.com/getarcaneapp/arcane/releases/tag/v1.13.2
github.comgithub.com/getarcaneapp/arcane/security/advisories/GHSA-2jv8-39rp-cqqr

Frequently asked

questions

Yes. Any attacker who can reach the Arcane manager's API endpoint can attempt to access remote environments via the affected /api/environments/{id}/... routes without providing credentials.

No valid credentials are required. Exploitation only needs network-level access to the Arcane API; authentication is bypassed because the proxy middleware forwards requests with a manager-held agent token.

The safest pattern is to enforce authentication and authorization before any routing or proxying decisions and before injecting internal credentials. All management and proxy endpoints should sit behind a consistent authentication middleware that cannot be skipped or reordered.

Adopt a security-by-design approach where:

  • All externally reachable APIs are authenticated by default.
  • Middleware ordering is explicitly reviewed so that authentication and authorization run first.
  • Internal tokens and service credentials are never attached to a request unless a verified user identity and appropriate permissions are present.
  • Automated tests and code review checklists validate that no management route is accessible without auth.
Find out what's already
exploitable in your codebase.