Severity:

CRITICAL RISK

(9.8)

Missing Authentication / Access Control in Yarbo Embedded

CVE-2026-7415

Analyzed

Vulnerability type

Category:

Missing Authentication / Access Control

CWE:

CWE-306

Impact:

Unauthenticated Remote Control and Data Exposure

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

The MQTT broker embedded in Yarbo firmware v2.3.9 is configured to accept anonymous connections and lacks topic-level access control.

Any device on the same network can subscribe to sensitive telemetry topics or publish control commands to the robot without any authentication or authorization, enabling full remote manipulation within network reach.

Affected context

This vulnerability affects Yarbo devices running firmware v2.3.9 that expose an embedded MQTT broker on a reachable network interface.

The issue becomes relevant when the robot is connected to a network segment where other hosts (wired or wireless) can reach the MQTT port, such as a home Wi‑Fi, corporate LAN, or shared IoT network.

Because the broker permits anonymous connections and does not enforce topic-level read/write permissions, there is no separation between trusted control clients and arbitrary network peers.

Exploit preconditions

  • Attack surface: MQTT service exposed by the Yarbo device on the local network (and potentially beyond if port-forwarded or reachable via VPN).
  • Required access: Network-level reachability to the device's MQTT port; no credentials, pairing, or prior trust relationship required.
  • Trigger point: Establishing an MQTT client connection and subscribing to or publishing on known or discoverable topics.
  • Why exploitation is possible: The broker is configured to:
  • Allow anonymous connections (no client authentication).
  • Provide unrestricted topic subscriptions and publications (no ACLs).
  • Treat all connected clients as fully trusted, regardless of origin.

Why this happens

The core issue is a lack of authentication and authorization controls on a network-exposed control protocol.

Instead of requiring credentials and enforcing per-topic access rules, the embedded MQTT broker is deployed in an open mode that implicitly trusts any client able to connect.

This aligns with CWE-306: Missing Authentication for Critical Function, because critical operations—robot control and access to sensitive telemetry—are exposed without any mechanism to verify the identity or privileges of the caller.

In effect, the MQTT endpoint is functioning as a critical control interface but is implemented as if it were a benign, read-only, or isolated interface.

Potential impact

  • Unauthenticated remote control of robot movement and behavior by any host on the same network.
  • Unauthorized access to telemetry data, including sensors, location, or other potentially sensitive environmental information.
  • Injection of malicious control messages that could cause physical damage to property or create safety risks for people nearby.
  • Disruption or denial of service by flooding the broker with control or subscription traffic.
  • Use of the robot as an entry point for broader network reconnaissance or as a pivot in multi-stage attacks.

General mitigation guidance

Mitigation generally focuses on enforcing strong identity and least-privilege access for all device control interfaces:

  • Require client authentication for MQTT connections, for example through credentials or certificate-based mechanisms appropriate to the product's lifecycle.
  • Introduce topic-level authorization rules so that only designated clients can publish to control topics or subscribe to sensitive telemetry.
  • Adopt secure-by-default configuration, shipping the device with closed or minimally exposed network services that must be explicitly enabled or paired.
  • Segregate device control traffic onto dedicated, least-privilege network segments, reducing exposure to untrusted devices.
  • Design the control plane so that safety-critical actions require explicit authorization and, where appropriate, local confirmation or bounded operating envelopes.

How teams detect issues like this

Detection can combine configuration assessment, network monitoring, and code/design review:

  • Review firmware configuration and code paths for MQTT broker setup, checking for anonymous access, missing authentication hooks, and absent ACL configuration.
  • Use static analysis or configuration scanning to flag embedded services started without authentication or with wildcard/allow-all access rules.
  • Capture and inspect network traffic to and from the device to confirm that MQTT sessions are established without any authentication handshake and that arbitrary topics can be accessed.
  • Incorporate security architecture reviews for IoT products, with explicit checks for unsecured control protocols and undocumented network services.
  • Employ automated firmware analysis tools that highlight open ports and default-open protocols as part of pre-release security validation.

Sources

SourceLink
github.comgithub.com/Bin4ry/yarbo-nat-in-my-back-yard
takeonme.orgtakeonme.org/gcves/GCVE-1337-2026-00000000000000000000000000000000000000000000000000111111111100111111111110000000000000000000000000000000000000000000000000000001001

Frequently asked

questions

It is directly exploitable from any network segment that can reach the MQTT port.

Whether this includes the wider internet depends on how the device is deployed—if the port is forwarded on a router, reachable via VPN, or placed on a routable segment, then internet-based attackers could exploit it as well.

An attacker only needs network-level connectivity to the device's MQTT service.

No user account, pairing, or prior interaction with the device is required, because the broker accepts anonymous connections and has no topic-level restrictions.

Treat the MQTT broker as a critical control interface: enforce client authentication, implement per-topic authorization rules, and ship the product with secure defaults that do not expose open control channels.

Additionally, constrain what actions can be triggered remotely and ensure sensitive operations have appropriate checks and safety limits.

Incorporate security into the architecture phase: define trust boundaries, require authenticated and authorized access for all control endpoints, and avoid "open by default" network services.

Regularly perform threat modeling, firmware configuration reviews, and network exposure testing as part of the development and release lifecycle to catch missing authentication and access control issues early.

Find out what's already
exploitable in your codebase.