Summary
- ENISA must publish all EUCC certificates, information on whether each is in force, suspended, withdrawn or expired, and the certification report corresponding to each one — an unusually public record for a security instrument.
- The highest label certifies that a product was “evaluated at a level intended to minimise the risk of state-of-the-art cyberattacks”, a statement about the depth of the examination, and the same scheme forbids the manufacturer self-assessment the parent regulation permits for low-risk products.
Europe's cybersecurity certification regime rests on two instruments: Regulation (EU) 2019/881, the Cybersecurity Act, which sets the framework and the three assurance labels, and Implementing Regulation (EU) 2024/482, which establishes the EUCC scheme that actually issues certificates for ICT products.
The report is part of the product. Article 42(1) of the EUCC regulation requires ENISA to publish “on the website referred to in Article 50(1) of Regulation (EU) 2019/881: (a) all EUCC certificates; (b) the information on the status of an EUCC certificate, notably whether it is in force, suspended, withdrawn, or expired; (c) certification reports corresponding to each EUCC certificate”. Article 10(5) closes the loop at the point of issue: “The certification body shall provide the national cybersecurity certification authority and ENISA with every EUCC certificate and every certification report in electronic form.”
That is the strongest public-record clause I have read in a European security instrument this month, and it is worth being precise about what it does not say.
The verb does the work. Article 52(7) of the Cybersecurity Act defines the highest label, and it should be read at the pace of the sentence: “A European cybersecurity certificate that refers to assurance level ‘high’ shall provide assurance that the ICT products, ICT services and ICT processes for which that certificate is issued meet the corresponding security requirements, including security functionalities, and that they have been evaluated at a level intended to minimise the risk of state-of-the-art cyberattacks carried out by actors with significant skills and resources.”
Evaluated at a level intended to minimise the risk. The assurance runs to the examination, not to the outcome. Nothing in that article tells a buyer that a high-assurance product withstands the attack it was tested against; it tells the buyer how hard someone was required to look, and at whom. The middle label is built the same way: ‘substantial’ covers “the known cybersecurity risks, and the risk of incidents and cyberattacks carried out by actors with limited skills and resources”, again as an evaluation intended to minimise them.
“High” is a claim about rigour. It is a good claim to be able to check. It is not a warranty.
Two labels, five levels, one refusal. Article 52(1) lets a scheme choose: “‘basic’, ‘substantial’ or ‘high’”. The EUCC scheme issues two. Article 4(1) to (3) is blunt about it: “Certification bodies shall issue EUCC certificates at assurance level ‘substantial’ or ‘high’. EUCC certificates at assurance level ‘substantial’ shall correspond to certificates that cover AVA_VAN level 1 or 2. EUCC certificates at assurance level ‘high’ shall correspond to certificates that cover AVA_VAN level 3, 4 or 5.”
Five AVA_VAN rungs, two labels: three under ‘high’, two under ‘substantial’. The reviewed text uses AVA_VAN without defining it, so this briefing will not pretend to explain what the metric measures. What the text settles is the commercial point — the cheapest label in the parent regulation is not on sale.
That label was cheap in a second sense too. Article 53(1) lets a scheme “allow for the conformity self-assessment under the sole responsibility of the manufacturer or provider”, but only “in relation to ICT products, ICT services and ICT processes that present a low risk corresponding to assurance level ‘basic’”: a self-declared sticker at the bottom of the range, and only if the scheme permits it, which Article 54(1)(e) requires every scheme to state.
The EUCC scheme says no, in full: “A conformity self-assessment within the meaning of Article 53 of Regulation (EU) 2019/881 shall not be permitted.” No carve-out for low-risk products. Whatever a buyer doubts about a certificate, the manufacturer did not sign it. That is the inversion at the centre of this instrument: the framework preserved the manufacturer's signature, the operational scheme deleted it, and the deletion sits in the same text that obliges the examiner's report to be published.
Who is allowed to look. Article 60(1): “The conformity assessment bodies shall be accredited by national accreditation bodies appointed pursuant to Regulation (EC) No 765/2008. Such accreditation shall be issued only where the conformity assessment body meets the requirements set out in the Annex to this Regulation.”
At the top level the framework adds a second gate. Article 56(6) provides that where a scheme requires ‘high’, the certificate “is to be issued only by a national cybersecurity certification authority or, in the following cases, by a conformity assessment body: (a) upon prior approval by the national cybersecurity certification authority for each individual European cybersecurity certificate issued by a conformity assessment body; or (b) on the basis of a general delegation of the task of issuing such European cybersecurity certificates to a conformity assessment body by the national cybersecurity certification authority.”
The EUCC regulation makes that gate concrete: a body must be authorised by the national authority to issue at ‘high’ in addition to accreditation (Article 21(1)), and an ITSEF needs the same authorisation to evaluate at ‘high’ (Article 22(1)). National authorities notify the Commission of the bodies competent to certify at each level (Article 23(1) and (2)) and send that notification to ENISA so that accurate eligibility information can be published (Article 23(4)).
One deadline here belongs to the Commission rather than to ENISA, and it is the only date of its kind in the reviewed text: “One year after the entry into force of a European cybersecurity certification scheme, the Commission shall publish a list of the conformity assessment bodies notified under that scheme in the Official Journal of the European Union.”
How long the paper lasts. Article 12 puts the clock and the escape hatch in one article. The certification body sets a validity period which “shall not exceed 5 years”; “by derogation from paragraph 2 that period may exceed 5 years, subject to the prior approval of the national cybersecurity certification authority” — and the authority must notify the European Cybersecurity Certification Group of the granted approval without undue delay. Protection-profile certificates run on a different clock, up to the lifetime of the profile concerned (Article 18).
That derogation is the clause to watch. It turns a ceiling into a decision someone has to take, record and defend.
Withdrawal is where the public record earns its keep. The issuing body must notify the national authority and ENISA of a withdrawal; the national authority notifies other relevant market surveillance authorities. On withdrawal the holder “shall disclose and register any publicly known and remediated vulnerability in the ICT product” on the European vulnerability database or the other repositories named in Article 55(1)(d) of the Cybersecurity Act. The certificate that leaves the shelf leaves a vulnerability notice behind it.
What I did not establish. I inspected no published EUCC certificate and no EUCC certification report. The publication duty is established from the articles quoted above; how it has been performed is not, and nothing here is a finding about ENISA's publication record. The agency's own certification topic page returned HTTP 429 during this research and was not read; the one ENISA sentence used here comes from the adjacent certification and standards page: “Certification is a tool allowing product vendors and service providers to demonstrate and advertise the cybersecurity of their solutions.”
Two further bounds travel with the facts. The EUCC quotations come from the original act as published, not from a consolidated version — ENISA reports later EUCC amendments dated December 2024 and December 2025, and those were not read. And the application date of the scheme, any transitional period, and any Union law making EUCC certification mandatory for a product category were not established from the reviewed text, so this briefing asserts neither that a certificate is voluntary nor that anyone must hold one. I read no penalty, sanction or enforcement action relating to certificates, and assert none.
What this adds to what we already carry. A search of this publication's archive for “ENISA” on 17 November 2026, limit 25, returned eight rows: seven published, one archived draft. Four concern the Cyber Resilience Act's manufacturer reporting route — three published pieces and one unpublished draft. Three concern adjacent instruments: incident-reporting routing, an NIS2 register proposal, the ATHENA and ENSOC threat-alert hubs. One concerns the European digital identity wallet. None of the three published reporting pieces returns a provision requiring the report itself to be published; a second search for “cybersecurity certification”, limit 20, returned the same set and nothing on the certification framework. The ENISA entry in the BTW directory carries the agency's card for Greece; I characterise only the eight rows that search returned.
Read the report, not the logo. Put the two halves together and the instrument is better than its reputation and weaker than its label. Better, because a stranger can check who was authorised, whether the paper is still in force, and what the examiner wrote. Weaker, because ‘high’ certifies the depth of an examination intended to minimise a named class of attack — and because the exemption a manufacturer might have wanted was written into the framework and then switched off in the scheme.
A procurement file that contains a certificate and no certification report has bought the label and skipped the part of this regime written for outsiders. The report is published. Read it.
Member Briefing
Read the full trend briefing
Members read the full briefing with the source notes behind every finding.
Strategic Circle
A daily briefing in your language, built from changes we have verified.
Join for USD 20/monthCancel anytime.
Members also get the full research library, a daily briefing on the networks they follow and the private forum.






