DEEBOT PRO CVE-2026-66411

The robot must be reachable over the network - attack vector is network (AV:N) per the CVSS vector, requiring no physical or local access. No authentication is required because the algorithm is incorrectly implemented, not merely absent, allowing an unauthenticated attacker to satisfy the handshake. No special configuration beyond network connectivity is needed; the flaw appears present in default device operation. Exploitation is constrained to the DEEBOT PRO M1 and K1VAC models specifically - no other Ecovacs products are confirmed affected. If the robot resides on a private LAN without external exposure, the attacker must first gain local network access (e.g., same Wi-Fi segment), which partially limits opportunistic exploitation from the public internet.

The CVSS 3.1 base score of 5.3 (Medium) reflects a network-accessible (AV:N), low-complexity (AC:L), unauthenticated (PR:N, UI:N) attack path with limited integrity impact (I:L) and no confidentiality or availability impact. The AC:L and PR:N combination lowers the exploitation bar significantly, requiring no special knowledge or prior credentials. However, real-world criticality is constrained by the nature of the affected asset - a consumer robot vacuum - which represents a low-criticality system in most environments. EPSS probability data is not included in available intelligence, so exploitation likelihood cannot be quantified numerically. This CVE is not in the CISA KEV catalog, and no public proof-of-concept has been identified, consistent with a coordinated disclosure via JVN/JPCERT rather than a discovered-in-the-wild scenario. The aggregate risk is further limited by the relatively narrow installed base of these two specific consumer models.

An attacker on the same local network - or reachable over the internet if the robot's WebSocket endpoint is externally exposed - initiates a WebSocket connection to the device. Due to the incorrectly implemented authentication algorithm (CWE-303), the handshake succeeds without valid credentials, granting the attacker an authorized-equivalent session from which they can issue commands to start or stop cleaning cycles, direct movement, or interact with any onboard sensors or camera features. No public proof-of-concept code has been identified at time of analysis.