Berlin (LabNews.ai) – The confirmation by the Federal Minister for Digital Affairs marks a turning point: For the first time since the Signal phishing wave in the spring, an attack on the mobile devices of members of parliament and government officials has been officially confirmed, one that is not based on human error, but rather on compromising the devices themselves. Karsten Wildberger (CDU) spoke on the ARD program "Bericht aus Berlin" about an Apple notification stating that "certain devices may have been compromised with so-called spyware." The Federal Office for Information Security (BSI) reacted "extremely quickly" and informed all affected users in writing.
The facts are currently limited, but reliable: Since October 29, iPhones have been receiving threat notifications from Apple warning of "Mercenary Spyware"—the category of commercial surveillance software that includes Pegasus from the NSO Group. The notification does not indicate whether the attack was successful; however, the BSI (Federal Office for Information Security) recommends that affected users "assume as a precaution that their device may be compromised." Wildberger did not provide details on the number or identity of those affected. An official attribution is still pending.
The Ring 0 question: Why kernel access is the real problem
The crucial technical unknown lies not in the spyware itself, but in the infection vector. Mercenary spyware campaigns against iPhones typically exploit chains of multiple vulnerabilities, including regular kernel exploits. The reason is structural: To operate spyware with the described capabilities—access to encrypted messaging apps, microphone, camera, and location—permanently and invisibly, the attacker must gain access from userland (Ring 3) to the kernel (Ring 0). Only there can security mechanisms such as code signing verification, sandboxing, and telemetry be reliably bypassed.
In the x86/x64 protection ring architecture, Ring 0 signifies absolute control: Whoever controls the kernel can read and write the address space of every userland process, bypass security callbacks, and falsify telemetry. The same logic applies to iOS, albeit in a modified form – the kernel manages memory, processes, and permissions. A successful kernel exploit is the gateway to persistent, undetected surveillance.
The documented Ring 0 escalation paths for Windows environments illustrate the problem area: kernel bugs in graphics drivers, file system filters, or VPN client drivers; BYOVD (Bring Your Own Vulnerable Driver) attacks, in which legitimate but vulnerable signed drivers are loaded to write to arbitrary kernel memory; and historically, compromised security products whose own drivers became the entry point. Whether such a chain was successful in the current case is not publicly known. The Apple warning itself does not distinguish between attempted and successful attacks.
Signal phishing as a contrasting example: What's different this time
The parallels to the spring are limited. In April 2026, attackers compromised Signal accounts through targeted social engineering – including those of Bundestag President Klöckner and Ministers Prien and Hubertz. The method was manual, not technical: attackers impersonated Signal support staff, stole security PINs, or persuaded targets to scan linking QR codes. The Federal Prosecutor's Office investigated on suspicion of espionage; the Federal Office for the Protection of the Constitution (BfV) attributed the campaign to a "state-controlled actor," with the chairman of the Parliamentary Control Committee blaming Russia.
The current incident differs fundamentally. Signal phishing compromises accounts, not devices. Attackers intercept what is transmitted through the service, but they do not gain control of the end device itself. Ring 0 spyware, on the other hand, can capture what happens before and after encryption: keystrokes, screen content, microphone, and camera. It is not limited to a single communication channel. For members of parliament and government officials, this means an expansion of the attack surface from the communication metadata level to the level of device integrity.
Neglected vulnerabilities beyond the zero-day narrative
Focusing on zero-day vulnerabilities obscures structural problems that are at least as relevant.
First: The signature and trust issue. As the analysis of the Berlin data theft showed, attackers were already misusing short-lived Microsoft Trusted Signing certificates in 2024/2025 to sign loaders as trustworthy. Antivirus software that relies on Authenticode reputation allows such signed binaries to pass through. This is not a classic code zero-day, but a zero-day in the trust model – and undermines recommendations such as "only allow signed software".
Secondly: BYOVD and the heterogeneity of government IT. Federal and state government agencies operate organically grown infrastructures with specialized hardware, legacy VPN clients, and specialized applications that require kernel drivers. HVCI (Hypervisor-Protected Code Integrity) and memory integrity are frequently disabled or in compatibility mode in such environments—not out of negligence, but because the applications wouldn't run otherwise. These very configurations open the door to BYOVD attacks.
Third: Fileless persistence and the limitations of signature recognition. Modern APT chains increasingly operate without persistent artifacts on the hard drive. PowerShell, WMI event subscriptions, scheduled tasks, and in-memory beacons leave no files that a signature-based scanner could find. The BSI recommendation to "only allow signed software" is ineffective here because the payload never exists as a file. Technically effective countermeasures—Constrained Language Mode, WDAC in Enforce mode, AMSI integration—are not consistently implemented across government IT infrastructure.
Fourth: The blind spot of mobile device management. While desktop IT is increasingly hardened through EDR and application allowlisting, politicians' mobile devices often remain outside of systematic management. Apple's threat notifications are a valuable early indicator – but they come from the manufacturer, not from an internal security infrastructure. Whether the affected devices were integrated into an MDM (Mobile Device Management) system that could have detected anomalies is not publicly known. The fact that the BSI (Federal Office for Information Security) was only alerted by Apple raises the question of whether the affected devices' own telemetry was sufficient.
Open questions and political implications
Three questions remain unanswered and are likely to gain importance in the coming days. First: Is this a targeted attack on a small number of high-profile targets or a broader campaign? Apple's warnings are sent to affected users, not to the public – the number of notifications is unknown. Second: Which actor is behind this? Mercenary spyware is typically licensed by commercial vendors to government clients. Without attribution, any identification remains speculative. Third: How many of those affected are using devices that are not integrated into a managed security architecture?
Wildberger expressed concern but avoided any speculation. The timing is politically sensitive: The German government's Digital Summit begins in Berlin on Monday, where Chancellor Merz and Wildberger will appear together. The fact that the confirmation of the attack comes immediately beforehand is likely to accelerate the political debate about the sovereignty and security of mobile communication.
For the affected members of parliament and government officials, there is an uncomfortable recommendation: Anyone who has received an Apple threat notification should treat their device as potentially compromised. This doesn't necessarily mean that data has been compromised – but it does mean that the device's integrity can no longer be assumed. In the logic of Ring 0 compromise, there is no partial security. Either the kernel is intact, or it isn't.
Note on the facts: This analysis is based exclusively on publicly available sources. Technical details regarding the infection vector, the number of affected individuals, and attribution are not yet publicly available. The Ring-0 analysis is based on documented attack patterns in comparable campaigns and general architectural principles, not on a technical investigation of the current incident.
