Cybersecurity Roboticsthe leading robot cybersecurity lab

Robots are the most critical cybersecurity frontier.

A compromised server leaks data. A compromised robot can kill — it welds, lifts, drives, or operates on a patient. Security flaws become safety risks in robotics. Below, we quantify how fast that frontier is widening, using the world's robot-vulnerability records.

≈ one new robot vulnerability disclosed every 4 days over the last three years — and the pace is rising.

547
robot vulnerabilities aggregated
RVD · NVD · EUVD · CISA KEV
298
rated high or critical severity
CVSS / RVSS
267
with an assigned CVE identifier
cve.org
16
landmark robot hacks (2017–2026)
see Milestones

The humanoid frontier

A commercial humanoid — the Unitree G1 — was shown to hand an attacker root access over Bluetooth, ship static cryptographic keys, and stream telemetry to overseas servers: a helper turned into a mobile surveillance and cyber-operations platform. This is where robot cybersecurity stops being theoretical.

The cybersecurity of a humanoid ↗   Humanoids as attack vectors ↗

A cyberattack that ends in the physical world

Robot intrusions follow the classic attack lifecycle — with two robotics twists. Many robots grant root on first access, so the privilege-escalation step simply vanishes; and the chain doesn't end in stolen data. It ends in actuation — a machine that welds, lifts or drives. That is why we treat every robot security flaw as a safety defect.

01ReconnaissanceROS / DDS discovery02Initial accessUSB · SSH · BLE pair03Lateral movementflat net · exposed PLC04Persistencefirmware · update chan.05Actuationmotor & force cmds06Physical harmwelds · lifts · drives↳ often root on first access
Stages after the Cyber Kill Chain and the ROS 2 threat model; robot techniques drawn from the Akerbeltz ransomware and humanoid case studies.

The robot's brain is now an attack surface

As robots hand control to foundation models, the attack surface moves into the reasoning loop. A manipulated image, a few injected words, or a poisoned policy no longer just leaks data — it becomes physical action. Jailbreaks like RoboPAIR drive language-controlled robots to unsafe behaviour; adversarial vision derails vision-language-action models; and a 1–5 word "safety" phrase can halt a robot outright — a semantic denial-of-service.

Perceptioncamera · LiDAR · micFoundation modelLLM / VLA reasoningActionmotor commandsPhysical worldadversarial visionprompt injection · jailbreakbackdoor · semantic DoS

Read the briefing: Foundation-model robots ↗

After SoK: Security & Privacy of Foundation-Model-Powered Robots (2026), RoboPAIR, and recent VLA / prompt-injection / semantic-DoS research. This is where robot cybersecurity is heading next.

Robot vulnerabilities over time

New disclosures per year (bars) and the cumulative total (line), aggregated across sources.

Sources: RVD, NVD (robot keywords) & EUVD — 547 records, deduplicated by CVE.

Severity distribution

Robot vulnerabilities by CVSS / RVSS severity band.

Severity from source metadata (NVD CVSS, RVD RVSS, EUVD base score).

Coverage by source

This lab does not rely on any single database — it fuses several, the way zerodayclock.com does for IT, but focused on robots.

RVD (github.com/aliasrobotics/rvd) · NVD (nvd.nist.gov) · EUVD (euvd.enisa.europa.eu) · CISA KEV cross-referenced.

Why robots are different

In IT, the worst case is data loss. In robotics, the worst case is a machine that hurts someone. Robots fuse IT, OT and IoT attack surfaces and add actuation, so the same bug class carries far more consequence — which is why we treat every robot security flaw as a safety defect.

ITOTIoT+ Actuationa body that movesPhysicalharmattack surfaces of any connected system……times a machine that welds, lifts, drives

See Reviewing the status of robot cybersecurity, Safety requires security in robotics, and the Milestones of how the threat escalated.

Explore all 547 vulnerabilities →

The window to patch is closing

Across all software, the median time from a vulnerability going public to its first exploitation in the wild has collapsed from 771 days in 2018 to same-day by 2025 — a 99.9% fall. A server can be patched inside that window; a fleet of certified, safety-critical robots — often offline for months, expensive to recertify — cannot. As exploitation turns instant, the machines that move inherit the sharpest end of it.

1000 d100 d10 d1 day201820192020202120222023202420252026771 days~5 dayssame-day
Median time-to-exploit, industry-wide: EUVD disclosure dates cross-referenced with CISA KEV exploitation records, 2018–2026; exponential fit R²=0.98. Motif after zerodayclock.com.

Most affected manufacturers

Which robot manufacturers appear most across the aggregated records — a rough map of where hardware risk concentrates. Middleware (ROS, DDS) is tracked separately.

Vendor inferred from RVD labels and NVD/EUVD descriptions; 377/547 records attributed.

Security is becoming a legal requirement

Regulators and standards bodies are catching up fast — and they now encode this lab's thesis. ISO 10218:2025 folds a mandatory cybersecurity risk assessment into robot functional safety, and the EU Cyber Resilience Act makes secure-by-design binding for any product with digital elements. In robotics, a security flaw is now — in law — a safety defect.

todayDec 2024CRA in forceApr 2025ISO 10218:2025Sep 2026CRA duties beginDec 2027CRA full effect
EU CRA (Reg. 2024/2847): obligations from 11 Sep 2026, full application 11 Dec 2027 · ISO 10218-1/-2:2025 (in force Apr 2025) references IEC TS 63074 → IEC 62443.
Research
Cybersecurity AI: Hacking Consumer Robots in the AI Era
Is robot cybersecurity broken by AI? Consumer robots -- from autonomous lawnmowers to powered exoskeletons and window cleaners -- are rapidly entering homes and workplaces, yet…
2026
Research
Foundation-model robots: a new attack surface
As robots hand control to large language and vision–language–action models, the reasoning loop itself becomes attackable. This briefing surveys the 2026 literature on…
2026
Research
Humanoid robots as attack vectors
We show that a commercial humanoid (Unitree G1) can be repurposed as a surveillance and cyber-operations platform: a mobile sensor suite with network access that an attacker can…
Sep 15, 2025
Research
The cybersecurity of a humanoid robot
As humanoids move from demos to deployments, we assess the cybersecurity of a commercial humanoid robot and find encryption flaws and unauthorized telemetry — showing that the…
Sep 14, 2025
Research
Offensive Robot Cybersecurity
Offensive Robot Cybersecurity introduces a groundbreaking approach by advocating for offensive security methods empowered by means of automation. It emphasizes the necessity of…
2025
Research
Robot cybersecurity, a review
Robots are quickly becoming the tools we deploy to act upon the world — yet they are being built with little regard for security. This review argues that in robotics, security is…
Jun 1, 2022
Research
Robot teardown: stripping industrial robots for good
Hardware teardown is a well-established practice in security research but under-applied to robots. This work argues that tearing robots down — physically and logically — is…
May 15, 2022
Research
SROS2: usable cybersecurity tools for ROS 2
Security in ROS 2 is only effective if developers actually use it. SROS2 provides a usable set of tools and a methodology to secure ROS 2 computational graphs, lowering the…
Apr 1, 2022
Archive
SROS2: Usable Cyber Security Tools for ROS 2
We propose a robot cybersecurity methodology to secure computational graphs and improve SROS2, usable security tools for ROS 2. We argue that without usability, security in robotics will be greatly impaired.
Mar 17, 2022
Archive
Robot Hacking Manual (RHM) v0.4
Learn robot cybersecurity through the Robot Hacking Manual (RHM), an introductory series about cybersecurity in robotics, with an attempt to provide comprehensive case studies and step-by-step tutorials.
Dec 13, 2021

Browse all research →