Skip to main content
Be Bitwise

Coordinated Vulnerability Disclosure

Responsible Disclosure & UK Law · Zero-Day Research track

7 min read

In this module

The Disclosure Spectrum

When you discover a vulnerability, you must decide how and when to share that information. This decision has been debated in the security community for decades, and three broad approaches have emerged:

Full Disclosure

The researcher publishes complete vulnerability details publicly — often on a mailing list like Full-Disclosure or a personal blog — without notifying the vendor first. The argument is that users deserve immediate information about risks, and that public pressure forces faster patching.

The obvious problem: attackers get the same information at the same time as defenders, before any patch exists. Full disclosure was more common in the early 2000s when vendors routinely ignored private reports. It's now a last resort, not a default posture.

No Disclosure (Secrecy)

The vulnerability is never reported. The researcher may keep it private, sell it to a government or broker, or simply choose not to act. This approach leaves all users vulnerable indefinitely - not only to the original discoverer but to anyone else who independently finds the same flaw.

Coordinated Vulnerability Disclosure (CVD)

The researcher notifies the vendor privately and allows a defined period for a fix to be developed and released before publishing details. This balances the vendor's need for time to patch with the public's right to know about risks.

Coordinated Vulnerability Disclosure

CVD (sometimes still called "responsible disclosure") is the accepted industry standard. The researcher contacts the vendor, provides technical details, and sets a deadline. The vendor develops a patch and coordinates a release date with the researcher. After the deadline - or after users have had an opportunity to apply the patch - the researcher publishes an advisory. If the vendor fails to act within the deadline, the researcher may publish regardless.

The 90-Day Standard

Google Project Zero, established in 2014, set the benchmark that most of the industry now follows. From 2021 until November 2026 its policy used a 90+30 model, which many other teams copied:

  1. The vendor receives 90 days from the date of notification to release a patch.
  2. If a patch ships within those 90 days, there's an additional 30-day period before full technical details are published, giving users time to apply it. If no patch ships, details go public at day 90.
  3. For actively exploited zero-days - vulnerabilities being exploited in the wild at the time of discovery - the deadline is just 7 days, reflecting the urgency when attackers are already causing harm.

Project Zero keeps changing the model. Since mid-2025 it has published the vendor, product and report date of each new report within about a week (its "Reporting Transparency" page), with no technical details. On 13 November 2026 it replaced 90+30 with a "Full X" policy: details are published X days after the report whether or not a fix exists, starting at 90 and with announced moves to 60 days in February 2027 and 30 days in January 2028. Policies like this shift, so check the vendor's or researcher's current published policy rather than relying on a summary.

The impact of fixed deadlines has been measurable. By 2019, 97.7% of bugs reported by Project Zero were fixed within 90 days, a dramatic improvement over rates before fixed deadlines existed. Hard deadlines, it turns out, concentrate minds.

Other Disclosure Timelines

Different organisations apply different timelines based on their assessment of the right balance between vendor time and user risk:

OrganisationStandard DeadlineNotes
Google Project Zero90 days, falling to 60 then 30 ("Full X", from Nov 2026)7 days for actively exploited zero-days; previously 90+30
CERT/CC45 daysHistorically one of the earliest CVD programmes
Zero Day Initiative (ZDI)120 daysOperates a vulnerability purchase programme
MicrosoftNo fixed public deadlineUses Coordinated Vulnerability Disclosure

There is no single "correct" timeline. The 90-day window has become the most widely cited standard, but the appropriate deadline can vary depending on the severity of the vulnerability, the complexity of the fix, and whether the vulnerability is being actively exploited.

Two international standards describe the process from the vendor's side: ISO/IEC 29147 covers receiving reports and publishing advisories, and ISO/IEC 30111 covers the internal handling between the two. Regulation is arriving too. Since 11 September 2026 the EU Cyber Resilience Act has required manufacturers of products with digital elements to notify ENISA and their national CSIRT of actively exploited vulnerabilities, starting with an early warning within 24 hours.

The Disclosure Timeline in Practice

Researcher
Vendor
Public
Coordinator
Day 0Day 30Day 60Day 90Day 120

Click a step to see details about the coordinated disclosure process

A typical coordinated disclosure proceeds through several stages. The researcher discovers and validates the vulnerability, then makes initial contact with the vendor's security team. Both parties agree on a timeline for remediation. The vendor develops, tests, and releases a patch. Finally, the researcher publishes an advisory crediting the discovery and providing technical details.

Communication throughout is non-negotiable. Vendors sometimes need extensions — a fix turns out harder than expected, or the vulnerability cuts across multiple products that all need patching simultaneously. Reasonable researchers accommodate genuine requests. Open-ended delays with no visible progress are a different matter.

The NCSC Vulnerability Disclosure Toolkit

The UK National Cyber Security Centre (NCSC) publishes a Vulnerability Disclosure Toolkit to help organisations establish their own disclosure policies. The toolkit provides:

  • Guidance on creating a vulnerability disclosure policy
  • Template language for communicating with researchers
  • Advice on setting up secure communication channels
  • Recommendations for internal triage and response processes

The NCSC actively encourages UK organisations to publish vulnerability disclosure policies so that researchers know how to report issues and what to expect in return. An organisation without a published policy creates uncertainty for researchers and risks losing reports entirely - or having details published without prior notice.

security.txt (RFC 9116)

RFC 9116, published in 2022, standardises a machine-readable file that websites can use to communicate their vulnerability disclosure policy. The file is named security.txt and is placed at /.well-known/security.txt.

text
# Example security.txt (RFC 9116)
Contact: mailto:security@example.com
Contact: https://example.com/security/report
Expires: 2027-01-01T00:00:00.000Z
Encryption: https://example.com/.well-known/pgp-key.txt
Policy: https://example.com/security/disclosure-policy
Preferred-Languages: en

Key fields include:

  • Contact (required): How to reach the security team - an email address, a web form URL, or a phone number.
  • Expires (required): When the file should be considered stale and re-fetched.
  • Encryption: A link to the organisation's PGP key for encrypted communication.
  • Policy: A link to the full vulnerability disclosure policy.
  • Preferred-Languages: The languages the security team can handle.
Finding the Right Contact

Before reporting a vulnerability, always check for a security.txt file at /.well-known/security.txt on the target's domain. If none exists, look for a security page on the website, check whether the vendor participates in a bug bounty programme, or search for a security@ email address. If all else fails, CERT/CC and national CSIRTs can act as intermediaries.

The existence of security.txt signals that an organisation has thought about disclosure and is prepared to receive reports. Its absence does not mean an organisation is hostile to research, but it does make the researcher's job harder.

When Coordination Breaks Down

Not every disclosure goes smoothly. Vendors may ignore reports, dispute the severity, threaten legal action, or simply take too long. When coordination fails, researchers face difficult choices.

The commonest failure is plain silence. Step through the timeline below for one way to handle a vendor that never replies: when to chase, when to hand the case to a coordinator, and whose clock you're on once you have.

When the vendor goes quiet

Researcher
Vendor
Public
Coordinator
Day 0Day 15Day 30Day 45Day 60

Day numbers are illustrative. The 45-day window is CERT/CC's published policy.

Click a step to see what to do when the vendor goes quiet

The community consensus, reflected in policies like Project Zero's, is that a fixed deadline provides the best balance. If the vendor has had 90 days and has not acted, publication serves the public interest by informing users of a risk they cannot otherwise assess. Where policies add a period after a patch ships, as 90+30 did, the aim is to keep technical details from being weaponised before defenders have had a chance to act.

CVD exists to protect the people actually affected by vulnerabilities — the users — as quickly as the process allows.