US Edition Austin, TX UTC 00:00:00

How to get a CVE ID: CNAs, the request form and what happens next

How to get a CVE ID: find the CNA covering your product, use GitHub or the MITRE CNA-LR form, and what the CVE record states mean. As of Oct 2026.

To get a CVE ID, you ask the CVE Numbering Authority (CNA) whose scope covers the affected product. That is usually the vendor or the open-source project's own CNA, or a platform such as GitHub for repositories hosted there. If no CNA covers the product, you can request an ID from the MITRE CNA of Last Resort (CNA-LR) through the CVE Program's request form. The process below reflects the MITRE and GitHub documentation as of October 2026; check the pages linked here for changes.

A CVE ID is only an identifier. It does not carry a severity score, a fix or an advisory. Those come from the CNA, the vendor and later enrichment by other parties.

What a CNA is and why it matters

A CNA is an organisation authorised by the CVE Program to assign CVE IDs and publish CVE Records for vulnerabilities within a defined scope. The Program describes the model on its CNA overview page. Scope is the key idea: a software vendor typically covers its own products, a research organisation may cover specific projects, and a hosting platform may cover the projects it hosts.

The practical rule is to go to the closest authority first. A vendor that is itself a CNA will usually be the quickest route for its own products, and it also controls the fix and the advisory, so the CVE record lines up with them. The Program's partner list on cve.org shows each CNA and its scope.

Step by step: how to request a CVE ID

  1. Confirm it is a vulnerability. Check that the issue is not already covered by an existing CVE for the same flaw. The CNA-LR accepts a request only if the claim is unique and valid under the CNA Operational Rules.
  2. Find the responsible CNA. Search the partner list for the vendor or project. Check the vendor's security page, which often says whether it is a CNA and how to report.
  3. Report privately first. If the vendor has not yet fixed the issue, follow its disclosure process. A security.txt file or a security policy page normally says where to send reports. Coordinated disclosure stays with you and the vendor; a CVE request does not hand that role to anyone else.
  4. Ask the CNA for an ID. Vendor CNAs usually assign the ID as part of triage. Give them what they need to judge the report.
  5. Fall back to the CNA-LR. If there is no CNA for the product, or the CNA cannot be reached or declines, use the MITRE request form.

What to prepare before you ask

Whoever assigns the ID needs the same core facts. The MITRE form lists these as the minimum:

  • the problem type, for example a CWE weakness class or the impact
  • the vendor or project name
  • the product name
  • the affected version or versions
  • a description of the vulnerability
  • a reference URL, such as an advisory, issue or commit

CVSS data is optional but helpful; our CVSS 4.0 guide explains how the metrics work. The form accepts CVSS 4.0 or 3.1.

Requesting through GitHub

For a public repository, the maintainer can request a CVE from inside a repository security advisory. According to the GitHub documentation, you first create a draft advisory and fill in the affected versions. Choosing to request the CVE ID later makes a Request CVE button appear at the bottom of the advisory in place of the publish button. GitHub presents the request as the route for when you do not already have a CVE ID for the vulnerability. The page gives no processing time, so plan for some delay before the advisory is published.

Requesting through the MITRE CNA-LR

The CNA-LR is a fallback, not the first stop. According to MITRE's CNA-LR page, it typically steps in when the product's supplier has no CNA, or when there is a valid dispute over another CNA's decision. MITRE calls a request from the supplier the preferred route; it accepts one from the discoverer if the supplier will not request an ID or cannot be contacted, and the claim is unique and valid.

The process on the request page works like this:

  1. Sign in with a Google account or a GitLab.com account.
  2. Choose "Reserve One CAN" to create a CAN tracking number.
  3. Enter the vulnerability details and, if you have it, CVSS data.
  4. Submit the request and wait for review.
  5. Check the CAN link over the following days for a message that the request was accepted and assigned a CVE ID.

A CAN number is not a CVE ID and should not be used publicly. MITRE gives no fixed response time, only that requests are reviewed as soon as possible, with acceptance expected "in the coming days", and that incomplete ones may be delayed. The CNA-LR also does not publish records for vulnerabilities that are not publicly disclosed, so expect an ID but not a public record if the flaw is still under embargo.

What happens after you get the ID

An ID that has been assigned but has no public details is reserved. It becomes visible once the CNA publishes the CVE Record. The CVE Record Format (CVE JSON 5) requires, in the CNA container, at least one English description, at least one affected product entry and at least one reference, per the schema documentation. A record is either published or rejected; a rejected record is one the CNA has withdrawn, for example as invalid or duplicate.

The official CVE List is open on GitHub in the cvelistV5 repository, which reports refreshing about every seven minutes. Records can also carry extra containers from other parties. CISA-ADP, for example, adds data such as SSVC and Known Exploited Vulnerabilities information. This enrichment is separate from assignment, so a new record may lack a score for a while.

If a CVE ID appears in a public advisory but the record is not yet on cve.org, MITRE offers a separate "Reserved But Public" form for that situation.

Common mistakes

  • Going to MITRE when a vendor CNA covers the product.
  • Quoting the CAN tracking number as if it were a CVE ID.
  • Submitting without version, impact or a reference, which delays review.
  • Requesting an ID for a duplicate of an existing CVE.

For news on newly published flaws, see our vulnerabilities coverage, and for wider context browse Security.

Frequently asked questions

How do I get a CVE number for a vulnerability I found?

Report it to the vendor or the CNA that covers the product and ask for an ID. If no CNA covers it, submit a request to the MITRE CNA of Last Resort.

Can I request a CVE for my own open-source project?

Yes. If the project is hosted on GitHub, a maintainer can request one from a repository security advisory. Otherwise look for a CNA that covers open-source projects, or use the MITRE form.

How long does it take to get a CVE ID?

There is no fixed time. MITRE and GitHub both publish no processing time, and vendor CNAs vary, so complete information helps avoid delays.

Is a CAN number the same as a CVE ID?

No. On the MITRE form, a CAN is a tracking number for your request. Use the CVE ID only after the request page says it has been accepted and assigned.

Does a CVE ID mean a vulnerability has a severity score?

No. Scoring and enrichment are separate from assignment and may be added later by the CNA or other parties.