← All industries

Type-approve the vehicle.
Ship the OTA. Survive the charger.

ISO/SAE 21434 CSMS conformance, UNECE WP.29 R155 and R156 type-approval support, TISAX assessor evidence, ISO 27001 alignment and NIS2 readiness for EU operations — packaged for OEM vehicle cybersecurity teams and the Tier-1 suppliers feeding them.

ISO/SAE 21434
CSMS lifecycle
R155 + R156
WP.29 type-approval
OCPP / ISO 15118
EV-charging surface
TISAX AL2 / AL3
Tier-1 supplier packs

Pain points the vehicle cyber lead and the homologation team argue about

CAN bus and gateway exposure

OBD-II remains the default attack foothold on every production vehicle. UDS (ISO 14229) Security Access 0x27 with short or default seed-and-key, routine-control 0x31 abuse, diagnostic-session 0x10 escalation, central-gateway misconfiguration leaking chassis CAN traffic into the infotainment domain — the canonical pattern across the Charlie Miller / Chris Valasek Jeep work, the Tesla Pwn2Own chains and the Keen Lab BMW research.

OTA update integrity

Manifest tampering, signing-key custody, downgrade and rollback abuse, partial-update interruption, A/B partition mishandling, secure-boot continuity across an OTA cycle. UNECE R156 makes this a type-approval question, not just an engineering preference.

EV charging station billing fraud

OCPP meter-value tampering, tariff manipulation, free-charge abuse via RemoteStartTransaction injection, RFID card cloning, ISO 15118 contract-certificate misbinding. The CPO eats the margin loss and the regulator eats the reputational hit.

Telematics and connected-car backend

Fleet-wide privilege escalation through the TSP back-end — single-tenant API flaws that translate into cross-vehicle command authority. The Sam Curry et al. 2023 disclosures against 16 OEM telematics back-ends remain the canonical case study for how a backend BOLA becomes a fleet-scale incident.

Tier-1 supplier risk

OEM Cybersecurity Interface Agreement flow-down under ISO/SAE 21434 Clause 7 forces Tier-1 suppliers to evidence CSMS-equivalent practice for every delivered component. Debug-port closure on production silicon, supply-chain integrity of the bootloader, TISAX AL2/AL3 assessor packs — all parked on the supplier's account manager.

NIS2 and DORA spillover into EU operations

NIS2 (Directive (EU) 2022/2555) brings large EU automotive operations into scope as essential or important entities. 24-hour early warning, 72-hour incident notification, management-board accountability. The cybersecurity organisation that grew around R155 now also has to answer to a NIS2 competent authority.

Compliance frameworks the engagement maps to

ISO/SAE 21434:2021 — Road vehicles — Cybersecurity engineering

link ↗

The international standard for road-vehicle cybersecurity engineering. Mandates a Cybersecurity Management System (CSMS) covering concept, product development, production, operation, maintenance and decommissioning. Clause 7 governs distributed cybersecurity activities and the Cybersecurity Interface Agreement (CIA) with suppliers. Clause 9 covers Threat Analysis and Risk Assessment (TARA). The technical evidence base UNECE R155 auditors expect.

UNECE WP.29 Regulation No. 155 (R155) — Cyber Security & CSMS

link ↗

Type-approval regulation for vehicle cybersecurity in UNECE 1958 Agreement contracting parties — the EU, UK, Japan, South Korea and ~60 others (not the US, China or India). Requires a valid CSMS certificate for the manufacturer plus per-vehicle-type cybersecurity assessment. In force for new vehicle types since 2022 and for all new vehicle registrations from July 2024 in the EU.

UNECE WP.29 Regulation No. 156 (R156) — Software Update & SUMS

link ↗

Companion regulation to R155 covering the Software Update Management System (SUMS) and the technical security of OTA updates — manifest integrity, signing chain, rollback prevention, RxSWIN (Regulation X Software Identification Number), driver notification and consent, abort-without-brick safety.

TISAX — Trusted Information Security Assessment Exchange

link ↗

VDA-derived assessment scheme governed by ENX Association. Information-security, prototype-protection and data-protection objectives at assessment levels AL1, AL2 or AL3. The audit scheme OEMs rely on to bind Tier-1 and Tier-2 suppliers to a common security baseline. AxVeil delivers the supplier-side evidence and pre-assessment gap-closure.

ISO/IEC 27001:2022

link ↗

The corporate ISMS baseline. Annex A control set updated in 2022 (93 controls, four themes: organisational, people, physical, technological). Sits beneath ISO/SAE 21434 as the broader information-security frame and is the contractual minimum many OEMs require from Tier-1 suppliers before TISAX is even considered.

NIS2 Directive (EU 2022/2555)

link ↗

Brings large EU automotive manufacturers and major operators of vehicle-related services into scope as essential or important entities. State-supervised cybersecurity risk-management obligations under Article 21, 24-hour early warning and 72-hour incident notification (Article 23), management-board accountability and personal liability for non-compliance.

ISO 15118 — Vehicle-to-Grid Communication Interface

link ↗

Plug & Charge protocol stack governing the EV-to-EVSE handshake, contract-certificate provisioning and the underlying V2G PKI. ISO 15118-2 (TLS-based) and ISO 15118-20 (extended for bidirectional and ACDP) define the secure-channel and authentication primitives that EV-charging pentests exercise end-to-end.

OCPP 2.0.1 (Open Charge Point Protocol) — Security Profiles

link ↗

Open Charge Alliance protocol for charger-to-back-end communication. Security Profile 3 (mutual TLS with client certificates) is the target end-state for new deployments. AxVeil testing covers authentication, message injection (RemoteStartTransaction, RemoteStopTransaction, ChangeConfiguration), firmware-update messages and tenant isolation across CPO back-ends.

Sample attack scenarios exercised

Three scenarios commonly run in an automotive engagement. Each draws on a public-record incident pattern or research disclosure and is reproduced on engineering mules, HIL benches or back-end staging environments — never on a customer vehicle or a production charger taking live transactions.

Scenario 1 — CAN injection via the OBD-II port
Physical OBD-II foothold against an engineering mule. UDS (ISO 14229) Security Access 0x27 seed-and-key analysis against a representative ECU — typical findings include short seed (16-bit) and key algorithms recoverable through ROM dump or RAM scrape. Diagnostic-session 0x10 escalation to programming session, routine-control 0x31 abuse against non-safety routines, central-gateway misconfiguration allowing chassis CAN frames to be injected from the infotainment-side OBD pin. Reproduces the access pattern documented in Miller & Valasek (2015), the Keen Lab BMW chain (2018) and the Tesla Pwn2Own work.
Scenario 2 — OTA spoofed firmware against the SUMS back-end
Tested against a staging copy of the OTA back-end and a bench update client. Manifest tampering against an unsigned or weakly-signed update bundle, downgrade attack swapping a current RxSWIN for an older known-vulnerable software identifier, partial-update interruption between A and B partitions to force the device into an inconsistent state, rollback-prevention bypass via NVRAM rollback. Maps each finding to the UNECE R156 SUMS requirement it threatens and to the relevant ISO/SAE 21434 Clause 9 TARA assumption.
Scenario 3 — EV charging billing fraud via OCPP and ISO 15118
Against a CPO staging environment with a representative DC fast charger model. OCPP 1.6-J fallback exploitation where TLS or security profile 3 is not enforced — RemoteStartTransaction injection from a spoofed back-end, MeterValues tampering to under-report energy delivered, free-charge abuse through ChangeConfiguration of the local-authorisation cache, RFID card cloning at the HMI. Where ISO 15118 Plug & Charge is enabled, contract-certificate misbinding tests against the V2G PKI. Outcome: a costed billing-fraud exposure model the CPO can hand to its finance and regulator-engagement teams.

Case study

Redacted reference — available under NDA

European OEM, single vehicle platform plus shared OTA back-end. 18-week engagement covering ISO/SAE 21434 CSMS gap assessment, R155 type-approval pre-assessment, R156 SUMS technical pentest against the OTA back-end and a HIL-bench client, and a CAN-bus and central-gateway pentest against the engineering mule. Findings: weak UDS Security Access on two ECUs, manifest-signing key custody concentrated in a single HSM with no documented rotation procedure, central-gateway misconfiguration leaking chassis CAN to the infotainment domain, partial-update abort path that left A and B partitions inconsistent under low-battery conditions.

Outcome: All Critical and High findings closed before R155 type-approval submission. CSMS gap-closure plan accepted by the OEM internal audit function. Tier-1 supplier evidence pack issued under the existing Cybersecurity Interface Agreement.

Full redacted report and reference call available under mutual NDA. Request via the scoping form →

Related work

Frequently asked questions

Do we actually need ISO/SAE 21434 if we already hold ISO 27001 and TISAX?+

Yes — they cover different scopes. ISO 27001 is the corporate information-security management system (ISMS) baseline. TISAX (Trusted Information Security Assessment Exchange, governed by ENX e.V.) is the VDA-derived audit scheme that OEMs use to bind their suppliers to a common information-security level — TISAX assessment levels AL1, AL2 or AL3, with information-security, prototype-protection and data-protection objectives. ISO/SAE 21434 is the standard for road-vehicle cybersecurity engineering: it mandates a Cybersecurity Management System (CSMS) covering the entire product lifecycle — concept, product development, post-development (production, operation, maintenance, decommissioning) — and is the technical evidence base that UNECE WP.29 R155 type-approval auditors expect. An OEM going to market in any of the 64 UNECE 1958 Agreement contracting parties (the EU, the UK, Japan, South Korea and most others, but not the US, China or India) cannot type-approve a new vehicle architecture in 2024 onward without a valid R155 CSMS certificate, and the technical substance of that certificate is ISO/SAE 21434 conformance. TISAX is the supplier-tier hygiene layer underneath. The three are complementary, not overlapping.

How do you pentest a CAN bus without bricking a vehicle or voiding type approval?+

Active CAN-bus testing is never performed on a vehicle that will be returned to a customer or used in further type-approval testing. The default rig is an engineering mule or a hardware-in-the-loop (HIL) bench supplied by the OEM's vehicle cybersecurity team, with a documented baseline ECU configuration that can be reflashed in minutes. Test surfaces include the OBD-II port (UDS / ISO 14229 diagnostic services, Security Access 0x27 brute force, routine control 0x31 abuse), the central gateway, the body and chassis CAN segments, and any exposed CAN-FD or automotive Ethernet (100/1000BASE-T1, SOME/IP) backbones. Findings are reproduced on bench, not on the road. Rules of Engagement explicitly exclude anything that could disable a safety-critical ECU (airbag, ESP, braking, steering, ADAS sensor fusion) on a vehicle that might return to public use. The deliverable supports the R155 type-approval cybersecurity assessment without exposing a production vehicle to destructive testing.

What does R156 add on top of R155, and how does that change OTA testing?+

UNECE WP.29 R155 covers the Cybersecurity Management System and the cybersecurity engineering of the vehicle type. UNECE WP.29 R156 covers the Software Update Management System (SUMS) and the technical security of the over-the-air update process itself — manifest integrity, package signing and chain of trust, rollback prevention, A/B partition handling, dependency tracking across ECUs, driver consent and notification, and the ability to abort a failed update without bricking the vehicle. R156 also defines RxSWIN (Regulation X Software Identification Number) so type-approval authorities can pin a specific software baseline to a specific vehicle homologation. OTA pentesting therefore exercises both the back-end update server (authentication, authorisation, tenant isolation across vehicle fleets, signing-key custody) and the in-vehicle update client (signature verification bypass, downgrade attacks, manifest tampering, partial-update interruption, secure-boot continuity). The deliverable maps each finding to the R156 SUMS requirement it threatens.

We operate EV charging stations as a CPO. Where does pentesting actually focus?+

Four surfaces. (1) The OCPP (Open Charge Point Protocol) link between charger and back-end — OCPP 1.6-J is still widely deployed but lacks transport security guarantees on its own; OCPP 2.0.1 adds security profile 3 (mutual TLS with client certificates) which is the target end-state. We test OCPP authentication, message-injection (RemoteStartTransaction, RemoteStopTransaction, ChangeConfiguration), tariff and meter-value tampering, firmware-update message abuse, and tenant isolation across CPO back-ends. (2) ISO 15118 — Plug & Charge — covering the EV-to-EVSE V2G_CI handshake, certificate provisioning, Mobility Operator (MO) contract handling, and the PKI underpinning all of it. (3) The HMI and payment surface — RFID and NFC card cloning, contactless EMV abuse, QR-code redirect, billing-fraud through transaction-meter manipulation. (4) The physical and serial-bus surface — exposed Ethernet ports, debug headers, J1939 / CAN-FD on DC fast chargers, and the supervisor-controller boundary. Findings tie back to the ENISA EV-charging threat landscape report and the NIST IR 8473 EV-extreme-fast-charging cybersecurity framework.

How does Tier-1 supplier risk fit into the engagement?+

OEMs flow ISO/SAE 21434 obligations down to their Tier-1 suppliers (and Tier-1s to Tier-2 component vendors) via a Cybersecurity Interface Agreement (CIA) — a contractual artefact required by Clause 7 of ISO/SAE 21434. The supplier is then expected to evidence CSMS-equivalent practice for its scope, typically via a TISAX assessment at the appropriate level (AL2 or AL3 for prototype data or higher-criticality components). AxVeil engagements for Tier-1 suppliers cover the technical pentest of the delivered ECU or system (UDS service surface, secure-boot chain, JTAG / SWD lockdown, debug-port closure on production silicon, supply-chain integrity of the bootloader and application binaries) and the evidence pack that the OEM's vehicle cybersecurity team needs to fold into its own R155 type-approval submission. We provide the CIA-aligned findings appendix so the supplier's account manager does not have to re-derive it under deadline pressure.

Scope an automotive engagement

Send the vehicle platform or component scope, the target regulatory milestone (R155 / R156 type approval, TISAX AL2 / AL3, ISO/SAE 21434 CSMS audit, NIS2 readiness), the bench setup available (engineering mule, HIL rig, staging OTA back-end, CPO test charger) and the OEM Cybersecurity Interface Agreement obligations in play. We respond with a fixed-fee proposal and a redacted reference under mutual NDA.

Request a scoping call