138% more reported CVEs per day in 2026 than last year.New: always-on Fixboxes. Vulnerable dependencies, fixed and verified.
← Back to Glossary

Vulnerability embargo

A vulnerability embargo is an agreement among the people who know about a vulnerability to keep it private until an agreed date, so fixes can be ready when it is made public.

Last updated: November 13, 2026

A vulnerability embargo is the period in coordinated vulnerability disclosure during which a vulnerability is known to a small group but kept from the public. Everyone in the group agrees not to publish, share or hint at the details before a set date and time, so that a fix, an advisory and often a CVE record can all go out together. The embargo ends at that date, or earlier if the details leak or the vulnerability is found being exploited.

Who is under an embargo

  • The reporter who found the vulnerability.
  • The vendor or maintainer who has to fix it, usually their security team or PSIRT.
  • Downstream vendors who ship the affected code and need time to build and test their own updates, such as Linux distributions or device makers.
  • Coordinators in multi-vendor cases: CERT/CC, a national CSIRT or a private list.

Only people who need to know are told, and each of them accepts the same terms. Customers and the public are not; that is the point.

Typical embargo lengths

There is no single standard. These are the published defaults of the best-known coordinators:

WhoDefaultExceptions
Google Project Zero90 days to ship a patch; details 30 days laterGrace period of 14 days on request (104 in total); 7 days when exploited in the wild
CERT/CCPublic 45 days after the initial report, whether or not a patch existsEarlier or later for active exploitation, especially serious or trivial issues, or changes to a standard
Zero Day Initiative120 days after the vendor respondsAbout 15 business days if the vendor never responds; shorter windows for incomplete patches
linux-distros listAt most 14 days; under 7 preferredThe reporter must post the details to oss-security when the embargo ends
Linux kernel security teamUp to 7 days from the start of the release process, once a fix existsUp to 14 days if the bug's criticality needs more time for release logistics

The numbers measure different things. Project Zero and CERT/CC count from the report and allow for the whole fix; the distribution lists count from when the details reach the list and assume a fix is nearly ready, so their embargo is only for packaging and release. That is why linux-distros asks reporters to tell upstream projects first and the kernel team asks reporters not to contact linux-distros until a fix has been accepted.

Example: an embargo on a distribution list

A maintainer of an open source compression library fixes a memory corruption bug and wants Linux distributions to ship the update on the same day as the upstream release.

  1. The maintainer prepares the patch privately and reserves a CVE ID.
  2. They email the linux-distros list with [vs] in the subject, PGP-encrypted, with the patch and a proposed public date five days later.
  3. Member distributions build and test their packages. They may share the details only with people on their own teams who need them, who accept the same terms. During the embargo the information is TLP:AMBER.
  4. On the agreed day the maintainer releases, the distributions push their updates, and the maintainer posts the full details to oss-security. The information is now public.

When an embargo breaks

An embargo can break because someone leaks it, because the fix lands in a public repository where others can read it, or because attackers are already exploiting it. When that happens, the usual answer is to publish what you have at once rather than keep the agreed date: the reason for waiting is gone, and defenders now know less than attackers.

The best-known case is Meltdown and Spectre. Project Zero reported the issues to Intel, AMD and ARM on 1 June 2017, and the coordinated disclosure was set for 9 January 2018. Google published on 3 January instead, "because of existing public reports and growing speculation in the press and security research community about the issue, which raises the risk of exploitation" (Google Security Blog).

The linux-distros policy states the obligations plainly: members must report a leak at once to the reporter and the list, investigate it and say how they will prevent it from happening again.

What to do as a downstream vendor

If you ship someone else's code and are told about a vulnerability under embargo:

  • Respect the terms. Share the details only with the people who need them, and do not commit the fix to a public branch before the date.
  • Find what is affected while the embargo lasts: which of your products and releases contain the vulnerable component, at which versions.
  • Prepare the update and the advisory so both are ready at the agreed time. What an advisory contains is covered in security advisory; machine-readable advisories use CSAF, and VEX tells customers which products are not affected.
  • Ask for more time early if you cannot make the date, not after it has passed.
  • Have a plan for a break: who decides to publish early, and what you publish if the fix is not ready yet, such as a mitigation.

Most downstream vendors are not told. They learn about the vulnerability when it is published, and then the time it takes to find the affected releases is time their customers stay exposed. Knowing exactly what each release contains shortens it.

How this relates to CRACI

CRACI does not take in vulnerability reports, manage embargoes or publish advisories. How to set up the process is covered in coordinated vulnerability disclosure, and the team that runs it in PSIRT.