Project Zero's Full X policy makes patch deployment part of the deadline
Project Zero has replaced its "90 + 30" disclosure policy with a staged "Full X" model. Under the old policy, a vendor had 90 days to remediate a vulnerability and could get another 30 days for patch rollout. The new policy publishes technical details X days after Project Zero reports the bug to the vendor, whether or not a fix has shipped. The current value of X is 90 days. It moves to 60 days on February 2, 2027 and to 30 days on January 3, 2028.
The change is explicitly tied to AI-assisted vulnerability discovery and exploitation. Project Zero says LLMs increase the speed of finding bugs and reverse engineering patches. It also says patch soak time is less useful when an attacker can derive bug details from a patch almost immediately. A period intended to protect users during rollout can create a false sense of secrecy while attackers use the patch as a map.
What Full X changes for defenders
The old timetable separated remediation from deployment. A team could argue that a fix existed before the deadline and then use a grace period to distribute it. Full X does not turn that argument into a safety property. It treats public technical detail as a date certain. The vendor's patch status may be tracked, but it is not a decision point that extends the embargo.
That creates a measurable patch gap: the time between a fix being available and the vulnerable user installing it. Project Zero says vendors need to reduce either that gap or the amount of time the vulnerable condition remains reachable. In practical terms, a security team needs an emergency deployment path for internet-facing systems and evidence that the path works.
The policy also removes most special cases, including the grace period and variant policy. The listed exceptions are active exploitation in the wild, details already made public, and a vendor request to disclose. The active-exploitation exception follows Google's application security policy, which includes a shorter seven-day disclosure path. Those exceptions are narrow. They should not be used as a reason to assume every report will be delayed.
Why severity is not the whole decision
Project Zero says it reports vulnerabilities with substantial severity and demonstrated security impact. It also says severity does not determine whether attackers will find a bug. A medium-severity defect can be valuable if it enables a chain, exposes a credential, or bypasses a policy that other systems depend on. A vendor that deprioritizes a report because of a label may still face disclosure at the scheduled time.
For defenders, the useful question is not "what is the CVSS score?" alone. Ask what the vulnerable component can reach, whether exploitation requires a user action, and whether the issue exposes identity material or a control-plane capability. Then map those answers to a deployment deadline. A bug in a reporting export may be less urgent than a bug in a credential manager, but the same ten-day response process cannot cover every case.
Operational changes that fit the policy
Maintain an inventory that can answer which systems run a named component and which owners can deploy a fix. Include containers, appliances, managed services, and dependencies that do not appear in a normal asset scan. Track the normalized patch time from release to verified production deployment. Record compensating controls separately when a fix cannot be deployed immediately.
Use staged rollout without allowing the last stage to become a permanent exemption. For high-risk services, test in a small environment, deploy to a canary, and keep rollback ready. For code with no safe patch, remove reachability with network policy, disable the vulnerable feature, or rotate the exposed credential. Project Zero's policy does not require a vendor patch for every possible mitigation, but it does require the defender to act before public details arrive.
The disclosure deadline is not a guarantee that an attacker waits for it. Automated exploitation can begin as soon as a patch diff is available, and some vulnerabilities are exploited before disclosure. Full X makes the defensive timeline clearer. It does not make disclosure safe by itself.
Project Zero's rationale is useful because it names the real bottleneck. Faster bug discovery is only one part of the problem. The other part is the gap between a vendor's build and a customer's deployed state. Security programs that measure patch availability but not patch adoption will look current while the exposed system remains old.
Top comments (0)