Google Project Zero has adopted a fixed disclosure clock: its default is to publish vulnerability details 90 days after reporting them to the vendor. Natalie Silvanovich’s November 13 announcement removes the separate 30-day patch-rollout period and most negotiated special cases.

The change affects how software and hardware vendors handle reports from this research team. It also matters to operators waiting for those fixes to reach packaged products, appliances and installed devices.

The 30-day deadline arrives in 2028

The announced schedule is:

Policy begins Default disclosure interval after vendor report
November 13, 2026 90 days, “Full 90”
February 2, 2027 60 days, “Full 60”
January 3, 2028 30 days, “Full 30”

Bill Toulas’s November 16 reporting describes the same staged transition. Thirty days is the announced destination, not today’s default.

Under the new rule, reaching agreement that a bug is fixed no longer determines the disclosure date. Project Zero says it will still track fixes and explain remaining exposure. Publication can arrive while a vulnerability remains unpatched.

Actively exploited vulnerabilities follow Google’s separate disclosure policy. Project Zero also retains exceptions for already-public details and vendor-requested publication. It says rare extended exceptions, such as those previously granted for Spectre and Meltdown, remain possible.

Measure protection reaching users

Project Zero argues that AI-assisted discovery increases the chance that different parties find the same flaw. It also says attackers can reconstruct vulnerabilities from released patches quickly enough to undermine a separate secrecy period after patch release.

Evidence status: an announced research-team policy and its threat assessment. This post identifies no particular newly exploited product, affected version or emergency patch. The policy is not a universal legal deadline for every vulnerability report.

The operational task is to shorten the patch gap, the interval between releasing a fix and installing it. For maintainers, that means planning who can validate a report, ship an emergency release and coordinate downstream packages. For operators, it means tracking whether the actual deployed product received the fix and whether activation requires a restart.

Project Zero’s earlier emergency-patching analysis distinguishes writing code from testing, delivery and activation. It discusses pretested feature switches, targeted filtering and alternate update channels. Those mechanisms need their own security review; a faster update route is also a privileged route for changing software.

The gradual schedule acknowledges that rushed fixes can harm users. Vendors also warned that a harsh deadline could pull effort away from more severe bugs reported elsewhere.

What the thread adds

In the Hacker News discussion, Aissen asks how the policy fits Android’s patch distribution chain. fc417fc802 replies that it should force that chain to speed up. These are a practical question and an expectation, rather than evidence that delivery has improved.

PeterStuer raises the burden on on-premises systems whose operators must integrate a component vendor’s fix. Wowfunhappy challenges the eventual 30-day interval: “What if the bug is actually inherently hard to fix?” The discussion exposes the tradeoff between faster disclosure and safely delivering a working repair.

The Android replies challenge the assumption that disclosure pressure reaches every vendor equally. sigmoid10 warns that budget-device suppliers may leave users exposed rather than accelerate updates. scott_w objects to expecting those users to buy their way out: “a low-income user has very little “wallet” to vote with.” These are warnings about who bears the risk, with no measured outcome for the new policy yet.

doodlesdev gives the opposing user-interest argument: “Then users might just as well know about it so they can avoid the software while it’s vulnerable.” Disclosure can inform a decision, but the replies leave open what happens when replacing or disconnecting the product is impractical.

From a 195-comment thread, in HN’s own order.

Caveat: Project Zero calls immediate exploitation after a patch a useful but imperfect attacker model. That is a planning assumption, not a measured claim that every released fix is instantly exploited.