Cyber Resilience Act · product security

Vulnerability management under the Cyber Resilience Act.

The CRA requires manufacturers to identify and document components and vulnerabilities, handle and remediate vulnerabilities during the support period, and report certain actively exploited vulnerabilities and severe incidents. Vellaci connects SBOM exposure, intelligence, human review, remediation, VEX and evidence in one product workflow.

Intelligence with provenance

CVSS, EPSS, known-exploitation signals (CISA KEV), fix availability, references and source feeds, each with the time it was fetched. KEV is a signal, never a legal determination.

Triage states

New, needs review, affected, not affected, mitigated, fixed, accepted risk, false positive, archived — with reasons required for not affected, accepted risk and false positive.

CRA review

Could this qualify as an actively exploited vulnerability under Article 14? Not reviewed → potentially reportable → confirmed or not reportable, with rationale, reviewer, timestamp and awareness time.

Remediation

Owner, deadline, fixed version, mitigation and ticket links (Jira, Linear, GitHub); tasks can be created from any finding and tracked to closure.

VEX

Assessments map to affected / not affected / fixed / under investigation with justification, ready for CycloneDX VEX export per product version.

Immutable history

Every status change, review and remediation update is appended to the decision history and the hash-chained audit log.

From alert to decision

What the CRA expects vulnerability handling to look like.

Annex I Part II of the Cyber Resilience Act lists the vulnerability-handling requirements manufacturers must meet for the whole support period: identify and document vulnerabilities and components (the SBOM), address and remediate them without delay, apply effective and regular tests and reviews, publicly disclose fixed vulnerabilities with advisories, enforce a coordinated vulnerability disclosure policy, provide a contact address, distribute security updates securely and free of charge. Vellaci turns each of those into a workflow with evidence:

Annex I Part IIRequirement (summary)Vellaci workflow
(1)Identify and document vulnerabilities and components, including an SBOMSBOM ingestion per version, deduplicated component catalogue, continuous matching
(2)Address and remediate vulnerabilities without delay, including security updatesTriage states, remediation owner and deadline, fixed-version tracking, release gates
(3)Apply effective and regular tests and reviewsReadiness controls with evidence and review dates; scan results attached as evidence
(4)Publicly disclose information about fixed vulnerabilitiesAdvisory drafts from the finding; VEX export; disclosure log
(5)Put in place and enforce a coordinated vulnerability disclosure policyCVD policy template, intake channel, SLAs, security.txt status
(6)Facilitate information sharing, including a contact addressPublished contact point, researcher acknowledgements, reporter communication log
(7)–(8)Secure update distribution; timely, free security updates with advisoriesUpdate mechanism documented per product; fix release linked to findings and advisories

FAQ

Triage questions, answered.

How is this different from a dependency scanner?
A scanner tells you a component has a CVE. Vellaci tells you which products and versions on the EU market ship it, whether exploitation evidence exists, who owns the decision, what was decided and why, whether Article 14 could apply, and what evidence supports all of that. Scanners produce alerts; manufacturers need decisions with names and timestamps.
What does 'priority' mean in Vellaci?
An explainable score combining CVSS severity, EPSS exploitation probability, CISA KEV listing, fix availability and your product's exposure context (network-exposed, safety-relevant, placed on the market). Every factor is shown next to the score; nothing is a black box, and the score never replaces the human reportability review.
Why are reasons mandatory for 'not affected' and 'accepted risk'?
Because those are the decisions an auditor, a customer or a market surveillance authority will ask about. A 'not affected' without a justification is a liability; with one — the vulnerable function is not reachable, the component is compiled out, a compensating control exists — it is a VEX statement and evidence of due diligence.
Does Vellaci decide what is 'actively exploited' under the CRA?
No. It surfaces evidence — a KEV listing, a high EPSS score, an exploitation report — and asks a named reviewer to record whether the vulnerability is potentially reportable, confirmed reportable or not reportable, with rationale and the awareness time. The regulation's definition (Article 3(42)) and your legal advisers govern the threshold.
Can we export VEX?
Yes. Assessments map to affected / not affected / fixed / under investigation with justification codes and are exported as CycloneDX VEX documents per product version, so downstream customers can consume your statements automatically.

Stop triaging in spreadsheets.

Connect exposure, decisions, deadlines and evidence.