Home/Services/Automotive & Embedded Penetration Testing
security service

Automotive & Embedded Penetration Testing

Automotive and embedded penetration testing across Europe: CAN bus, ECUs, firmware and hardware tested to ISO 21434 and WP.29. Fixed quote, free retest.

Manual, expert-ledEvidence-based findingsFree remediation retest

Automotive and embedded penetration testing is hardware-level offensive work: we attack ECUs, the CAN bus, telematics units, infotainment and the firmware inside them the way a determined adversary with the physical device would. When a vulnerability can affect a moving vehicle or a deployed fleet, testing has to reach the silicon, not stop at the network.

An embedded device is a computer you cannot patch on a whim, often sitting in a hostile physical environment for a decade. In vehicles the stakes are higher still, because a flaw can cross from an entertainment feature to a safety-critical function if the internal separation is weak. We test the whole stack, from the radio interface down to the flash chip, and we do it against the standards regulators now require.

What automotive and embedded penetration testing covers

We treat the device as an attacker with hands on it would: extracting firmware, listening to and injecting on internal buses, abusing debug interfaces, and attacking the wireless surfaces that let an adversary reach the device without touching it. The goal is to find the paths that lead from an accessible interface to control of a function that matters.

CAN bus, CAN FD and Automotive Ethernet traffic analysis and message injection
ECU and microcontroller firmware extraction, analysis and reverse engineering
Debug interface attacks over JTAG, SWD and UART, and flash dumping via SPI or I2C
Keyless entry and immobiliser testing, including relay and replay attacks
Telematics, infotainment and connected-vehicle backend and API testing
Secure boot, firmware signing and update mechanism review
Wireless surfaces: Bluetooth, BLE, Wi-Fi, cellular and RF key fobs

How we test an automotive or embedded target

The work is a mix of hardware bench analysis and software reverse engineering, informed by threat modelling to ISO/SAE 21434. We start by understanding the device’s architecture and its trust boundaries, then attack the interfaces in order of how an adversary would actually reach them.

Threat analysis and risk assessment

ISO/SAE 21434 frames automotive security around a TARA, a threat analysis and risk assessment. We build one for the target so the testing effort concentrates on the assets and attack paths that carry real safety and security consequence, not evenly across every pin on the board.

Hardware analysis

On the bench we identify the components, locate debug and test points, and try to gain a foothold: dumping firmware from flash over SPI, opening a UART console, or reaching a live JTAG or SWD interface. A surprising number of production devices ship with debug access still enabled and readout protection never set.

Fault injection and side-channel

Where secure boot or a locked interface stands in the way, we assess resistance to hardware attacks such as voltage glitching and, where relevant, side-channel analysis against cryptographic operations. These are the techniques a serious attacker uses to defeat a lock the software assumes is unbreakable.

Firmware reverse engineering

With firmware in hand we analyse it: unpacking images with binwalk, disassembling with Ghidra, and hunting for hardcoded keys, backdoor accounts, unsafe update logic, and the classic memory-corruption bugs that embedded C is prone to. Firmware is where the device’s real secrets and its real weaknesses live.

In-vehicle and bus testing

For automotive targets we work on the CAN bus and Automotive Ethernet, capturing and replaying traffic, fuzzing UDS diagnostic services under ISO 14229, and testing whether a message from a low-trust domain such as infotainment can influence a higher-trust domain. Domain isolation is the property that keeps an entertainment bug from becoming a braking problem, and we test it directly.

Wireless and backend

We test the remote attack surface too: Bluetooth and BLE pairing and profiles, Wi-Fi, the cellular telematics path, RF key fobs, and the cloud backend and APIs that a connected vehicle or device talks to. A modern car is also a fleet of internet endpoints, and those get tested to web and API standards as well.

Bench
hardware-level testing, not just network
21434
testing aligned to the automotive cybersecurity standard
Free
retest once you have fixed

Vulnerabilities we commonly find

Embedded and automotive systems share a set of recurring weaknesses, most of them rooted in the assumption that physical access is hard and therefore does not need defending against.

Open debug interfaces and unprotected flash

JTAG, SWD or UART left enabled in production, and flash memory without readout protection, so the entire firmware and its secrets can be extracted on a bench in minutes. This is the most common foothold we find.

Hardcoded secrets and weak keys

Signing keys, API tokens and diagnostic passwords baked into firmware and shared across an entire product line, so extracting one device compromises them all. Embedded crypto also tends to lean on weak or predictable random number sources.

Missing or bypassable secure boot

Devices that will run unsigned firmware, or whose secure boot can be defeated with a glitch, letting an attacker install persistent malicious code that survives updates.

Weak internal isolation

In vehicles, a CAN architecture where a compromised infotainment or telematics unit can inject messages onto buses controlling more sensitive functions. This is the flaw that turns a convenience feature into a safety issue.

Insecure update mechanisms

Over-the-air or diagnostic update paths without proper signature checks or rollback protection, which let an attacker push malicious firmware or downgrade to a known-vulnerable version.

Tools and equipment

The work uses logic analysers and oscilloscopes for signal analysis, hardware tools such as the Bus Pirate and flash programmers for memory access, CAN interfaces and SocketCAN with tooling like can-utils for bus work, software-defined radio for RF, and Ghidra with binwalk for firmware reverse engineering. Fault-injection rigs are used where secure boot needs to be assessed. The equipment gets us onto the device; the findings come from an engineer who understands how these systems fail.

Get a fixed quote

Want this tested on your own systems?

Free 20-minute scoping call, a fixed price with no hourly surprises, and a free retest once you fix what we find.

Get a fixed quote

Standards and compliance

Automotive security is now regulated, not optional. ISO/SAE 21434 defines cybersecurity engineering across the vehicle lifecycle, and UNECE WP.29 regulation R155 requires a certified cybersecurity management system for vehicle type approval, with R156 covering software updates. Our testing and reporting are structured to feed the evidence these require, including the TARA and verification activities. For general embedded and IoT devices we work to the ETSI EN 303 645 baseline and OWASP IoT guidance, and align to ISO 27001 where it applies. We provide the attestation and evidence mapped to your target framework.

What you get

The deliverables are written for both your embedded engineers and the people answering to a regulator or type-approval authority.

  • An executive summary framing the safety and security risk in terms leadership and auditors can act on.
  • A technical report with each finding scored, the attack described in reproducible detail with the hardware setup used, and a concrete fix.
  • The threat analysis and testing evidence structured to support ISO/SAE 21434 and WP.29 activities.
  • Prioritised remediation, a call with the reviewing engineer, and a free retest after you fix.
  • An attestation letter mapped to your compliance obligation where you need one.

How a remote attack on a vehicle chains together

The attacks that make headlines are not single bugs, they are chains, and understanding the chain is why we test end to end. It often begins at a wireless surface an attacker can reach without touching the car: the cellular telematics link, Bluetooth, or a companion app and its backend. A weakness there gets the attacker onto the telematics or infotainment unit.

Crossing from convenience to control

The dangerous step is the next one: moving from that low-trust unit onto the internal buses carrying messages to systems that affect how the car behaves. If the domains are properly separated, the attacker is stuck with an entertainment bug. If they are not, an infotainment compromise reaches far more than it should. Testing whether that boundary holds is the core safety question, and we probe it directly.

The backend multiplies the risk

Where a flaw sits in the connected-vehicle backend rather than the car, the impact scales, because one server talks to an entire fleet. A backend authorisation flaw that lets one account send commands to another owner’s vehicle is a fleet-wide problem, so we test those APIs with the same rigour as the hardware.

Testing safely, without bricking your hardware

Hardware testing carries real risk to the device, and we manage it deliberately. We work on sample units you provide rather than production stock, keep known-good firmware images so a device can be restored, and agree in advance which destructive techniques, such as fault injection that may damage a chip, are in scope. For in-vehicle work we prefer a bench setup or a dedicated test vehicle over anything a customer will drive. If a sample device is damaged during an authorised hardware attack, that is a result worth having on a test bench rather than a surprise discovered by an attacker in the field.

Pricing

The automotive embedded penetration testing cost depends on the number of components in scope, whether the work is firmware-only or full hardware bench analysis, and how much of the wireless and backend surface is included. Hardware testing takes time, and we scope it honestly.

Engagement What’s included Timeline Price
Single device / ECU One embedded device or ECU, firmware extraction and analysis, debug-interface and basic hardware testing, report and free retest 5–8 working days from €4,000
Device + wireless Full hardware bench analysis, firmware reverse engineering, wireless surface (BLE, Wi-Fi, RF), update-mechanism review, exec + technical report 8–12 working days €6,000–€15,000
Automotive system Multiple ECUs, CAN and Automotive Ethernet, UDS diagnostics, domain-isolation testing, telematics backend, TARA aligned to ISO 21434 12–20 working days €15,000–€40,000
Compliance add-on Evidence pack and attestation for ISO/SAE 21434 or UNECE WP.29 R155/R156 with any tier from €1,500
Custom / whole vehicle Full vehicle or a product platform of many devices, scoped to your needs on scoping custom

Every engagement is fixed-price, quoted after a free 20-minute scoping call — no hourly surprises, and a retest is included. Get a fixed quote

FAQ

How much does automotive and embedded penetration testing cost?
It starts from €4,000 for a single device or ECU and scales with the number of components and whether the work is firmware-only or full hardware bench analysis. You get a fixed quote after a free scoping call, with no hourly billing.
Do you need physical devices to test?
For hardware and firmware work, yes, usually one or two sample units we can open and instrument on the bench. Some firmware and backend analysis can be done from images and access alone, and we agree the sample requirements during scoping.
Do you test to ISO 21434 and UNECE WP.29?
Yes. We build the threat analysis and risk assessment ISO/SAE 21434 expects and structure the testing and evidence to support WP.29 R155 and R156 activities toward type approval.
Can you extract and reverse engineer firmware?
Yes. We dump firmware over interfaces like SPI, JTAG and UART where they are accessible, then analyse it with binwalk and Ghidra to find hardcoded secrets, unsafe update logic and memory-corruption bugs.
Do you test keyless entry and relay attacks?
Yes. We test remote keyless entry, immobilisers and RF key fobs, including relay and replay attacks, using software-defined radio and the relevant RF tooling.
What about the connected-car backend and app?
We test those too. The telematics cloud, its APIs and any companion mobile app are part of the modern vehicle attack surface, and we assess them to web, API and mobile standards alongside the hardware.
What will you actually deliver?
An executive summary, a technical report with each finding scored and reproducible including the hardware setup, evidence structured for ISO 21434 and WP.29, and a free retest after you fix.
Is a retest included after we fix?
Yes, and it is free. Once you have addressed the findings we re-test the affected components and confirm the fixes hold before updating the report and attestation.

Related services

Who needs this

Vehicle manufacturers and tier-one suppliers pursuing UNECE WP.29 type approval, teams building ECUs, telematics units or connected-vehicle platforms, and companies shipping embedded or IoT hardware that must be secure in the field. We are a European offensive-security team working with clients EU-wide and remotely.

why safetybis

Security you can prove

The same standard on every engagement, big or small.

Evidence, not opinions

Every finding ships with a reproduction and proof of concept — no vague "maybe vulnerable".

Humans over scanners

Certified engineers find the logic flaws and chained attacks automated tools walk straight past.

Fixed price, free retest

You know the cost up front, and verifying the fix is part of the deal — not a second invoice.

500+
assessments delivered
<30min
incident first response
12k+
infections removed
98%
fixed within one retest
$ safetybis quote --service "Automotive & Embedded Penetration Testing"

Ready to lock this down?

Free scoping call, fixed price, free retest. Tell us what you're running and we'll take it from there — usually within one business day.

get in touch

Tell us what you're running

Scoping is free. We reply within one business day, and under 30 minutes for active incidents.