Autel MaxiCharger CVE-2026-8986

Exploitation requires the attacker to control the OCPP server side of the charger's connection - either by operating a malicious OCPP backend, compromising the legitimate charge-point management server, or man-in-the-middling an OCPP session that lacks TLS/mutual authentication - and then sending a GetDiagnostics request with a crafted diagnostics URL containing OS command-injection payload. This is the AT:P (attack requirement) reflected in the CVSS 4.0 vector: it is not exploitable by an arbitrary unauthenticated internet host against a default charger unless that host can pose as the trusted OCPP server. Limiting factors: the charger must be running firmware V1.03.51 or earlier, must accept and act on GetDiagnostics from the connected server, and the attacker must already occupy the trusted-server role; properly authenticated/encrypted OCPP links and network isolation substantially raise the bar. No end-user interaction on the charger is needed (UI:N) and no credentials on the charger itself are required (PR:N).

The CVSS 4.0 base score is 9.5 (AV:N/AC:L/AT:P/PR:N/UI:N with high impact to both the vulnerable system and subsequent systems), reflecting network reachability, low complexity, and total compromise - but the AT:P (attack requirement present) metric is the crux: exploitation is not against an arbitrary anonymous attacker but requires the adversary to occupy the position of the charger's OCPP server, either by operating a malicious backend, compromising the legitimate one, or performing a man-in-the-middle on an unauthenticated/weakly-authenticated OCPP session. Publicly available exploit code raises practical risk, but this is a POC signal, not confirmed active exploitation - there is no CISA KEV listing and no EPSS score in the provided data, so widespread automated exploitation cannot be asserted. Net assessment: a genuine high-priority issue for operators whose chargers connect to OCPP servers over untrusted networks or without TLS/mutual authentication, but the 'malicious/compromised server' precondition meaningfully narrows the realistic attacker set compared to a raw 9.5 read.

An attacker who stands up a rogue OCPP server, compromises the legitimate charge-point management platform, or intercepts an unauthenticated OCPP session sends a GetDiagnostics request whose diagnostics URL embeds shell metacharacters and an attacker command. The charger parses the request and executes the injected command on its OS, giving the attacker code execution on the charging station; a public POC from CyberDanube demonstrates the technique, and the low attack complexity means it can be triggered with a single crafted message once server position is achieved.