Signing passport data, and the gap a signature does not close

What is actually being standardised

The current European DPP standards work does include a specific project for authentication and trusted data exchange. EN 18246 is scoped around the integrity, authenticity and reliability of DPP data exchange, including a framework for secure processing and communication across multiple actors.

That is relevant to fraud. Better verification of who produced a record, whether it has been altered and whether the exchange can be trusted can reduce opportunities for fraudulent or counterfeit data to move through the system.

The standard still does not close the physical-object gap. Its published project scope excludes DPP architecture itself, secure elements associated with data carriers and cryptographic security features for unique product identifiers. In other words, it can harden trust in the data exchange without defining the secure binding between that record and the physical item.

The project record for EN 18246 was at stage 50.60 from 16 July 2026 in the official standards-project material reviewed for this page. It is not treated here as OJEU-cited, so this page does not claim that it currently carries a legal presumption of conformity under ESPR. The same discipline applies to EN 18239 on DPP access rights: an EN project stage, national adoption record and OJEU citation are different events.

That is the useful standards boundary. Data authentication is becoming more specified. The question “is this the real physical product?” still needs its own mechanism.

For that physical boundary, see identity, authentication and ownership. For the legal standards status, use the DPP standards status owner.

What a signature does close

Worth being clear, because the gap this page describes is not an argument against signing.

Origin. A reader can establish which operator published a value rather than inferring it from the domain the page happens to sit on.

Alteration. A value changed after publication stops verifying. That is a real protection against a record being edited in transit or at rest by somebody other than the operator.

Attribution over time. A signed value carries who said it, which is exactly the discipline the estate already argues for on its own terms. Who inside a business owns a published field, and what happens when one turns out to be wrong, is a governance question with its own page, and a signature makes the answer checkable from outside rather than merely recorded inside.

Those three are not small. In a category where most systems export values without provenance, cryptographic attribution of origin is a meaningful improvement, and what a usable export has to contain is set out on what you can take with you when you leave a provider.

What a signature does not close

It does not bind the record to the object. This is the load-bearing point. The link between a physical garment and a signed record is a carrier: a printed code, a tag, a chip. Signing the record at the far end of that link does nothing to the link itself. Photograph a legitimate code, reprint it and every counterfeit resolves to a genuine, correctly signed passport.

The published data carrier standard confirms this from the other direction. Its scope covers symbology, encoding, print and production quality, durability and placement, and it explicitly excludes secure elements and any other cryptographic security features. The standard that governs the physical link says, in its own scope, that binding and secrecy are not its job.

So the two standards between them cover the record and the mark, and neither covers the join.

It does not make a value true. A signed statement that a garment is 60 per cent recycled polyester proves the operator said it. Whether the operator could substantiate it is a different question with its own tests, and what a passport field can and cannot prove is a separate discipline.

It does not confer a presumption. Not today. The reference is not published, so conformance with it buys nothing under the framework, however good the engineering is.

Where a verifiable credential adds more than a signature

A verifiable credential can put the signature or proof inside a wider machine-readable trust structure. Alongside the cryptographic proof, a verifier can inspect issuer identifiers and, where used, credential-status information such as suspension or revocation.

That helps answer more precise questions: who controlled the issuing identifier, has the credential changed and is it still current under its status mechanism?

It still does not answer the hardest questions automatically. The credential can contain a false claim. The issuer may not have the authority or competence the relying party expects. The evidence may not apply to the product. The signed record may still have no secure physical binding to the object carrying the code.

So the right progression is not “signature, then truth”. It is signature/proof → issuer and status checks → evidence and authority → physical binding. Each layer closes a different gap.

See what a verifiable credential actually proves for the full five-layer model.

Implementation examples are choices, not mandates

Current pilot software shows how these layers can be assembled without proving that one technical stack is legally required. CIRPASS-2's mock EU Registry, for example, can verify detached JWS signatures and issue signed proofs, but the project explicitly presents the implementation as exploratory and non-normative.

That makes it useful implementation evidence, not a mandate. It would be wrong to infer from one pilot that EN 18246 or EU DPP law requires JWS, one credential format or one signing technology.

What actually authenticates an object

Unchanged by any of this, and worth restating because the standards work invites the wrong conclusion.

Binding a record to a specific physical item requires something that lives in the object and is hard to copy. That is the reason every serious operator in the anti-counterfeiting market sells a physical or covert feature alongside the identifier rather than relying on the code, and it is why the party with the strongest commercial incentive to replace inspection with a lookup still pays people to inspect.

A signature is a good answer to "is this record genuine". It is not an answer to "is this object genuine", and the two questions have different suppliers.

Sources

• FprEN 18246, data authentication and integrity for digital product passports
  European standard at formal vote

  Subject and citation state reached through two independent secondary readings. Neither the draft nor its requirements were read, and nothing on this page describes a requirement inside it. Used for two propositions only: that the subject is being standardised as electronically signed data constructs, and that its reference is not published in the Official Journal.

• FprEN 18239, access rights management
  European standard in final draft

  Subject and citation state reached the same way, and carried here only as the second of the two standards outside the presumption.

• BS EN 18220:2026, Digital product passport. Data carriers
  European standard adopted as a British Standard, published 31 May 2026

  Scope read at the national standards body that publishes it. The load bearing source on this page, cited for its stated exclusions: architecture and use cases, secure elements, and any other cryptographic security features.

• Commission Implementing Decision (EU) 2026/1736 of 14 July 2026
  In force

  Used for which references are published and therefore which standards carry a presumption. Confirmed at the Commission's harmonised standards page and through a testing body's account of the decision.

Sources as at 30 August 2026.