Skip to content
AI.info

The Pulse

Anthropic cuts vulnerability disclosure windows to 30 and 45 days

Anthropic will shorten its disclosure window for high- and critical-severity vulnerabilities from 90 days to 30 days. The changes apply to reports sent to vendors and maintainers from January 1, 2027, while existing reports keep their deadlines.

Anthropic cuts vulnerability disclosure windows to 30 and 45 days

High- and critical-severity software vulnerabilities reported by Anthropic will generally face a 30-day disclosure window, down from 90 days, while other issues will move to a 45-day window. In a November 16 post, the company said the change reflects its view that AI is making it more likely that the same flaw will be found independently.

The new windows apply to vulnerabilities reported to vendors and maintainers on or after January 1, 2027. Reports already in progress on that date will keep their existing deadlines.

Shorter waits for escalation and details

The company will aim to escalate a finding to an external vulnerability coordinator after 15 days, rather than 30, if a maintainer does not respond to the initial report. It will also wait 30 days, rather than 45, after a patch becomes available before sharing full technical details.

The company's existing policy says the wait after a patch is intended to give downstream users time to deploy a fix before detailed exploitation information becomes public. The policy applies to vulnerabilities Anthropic discovers in open-source software and in closed-source software where it has authorization to conduct security research. It generally does not cover vulnerabilities reported to Anthropic by external researchers.

The company is also notifying maintainers of high- and critical-severity vulnerabilities that were reported more than 90 days ago, have passed the current policy deadline and have not yet been publicly disclosed. Those maintainers will receive an additional 30 days, after which Anthropic will publish the reports on its dashboard under the updated policy, starting January 1, 2027.

Exceptions for difficult fixes

Some vulnerabilities cannot always be resolved within the new windows, including flaws in software that is hard to deploy, such as critical infrastructure, and those requiring changes to a specification or RFC. In those cases, Anthropic will work with the vendor on a remediation timeline.

The seven-day disclosure window for any vulnerability under active exploitation remains unchanged. The shorter timelines will not apply to AI-discovered findings that have not received human review, including findings from OSS Scanner.

Anthropic said that frontier models have become better at finding vulnerabilities over recent months. It said it is seeing more independent rediscovery of vulnerabilities as those capabilities improve, and that models can also help develop fixes.

Reporting and maintainer support

The company will continue to include actionable guidance in each report and, where possible, a patch or mitigation. It will also continue to agree on a sustainable pace with maintainers before sending a large volume of findings to a single project.

Anthropic's dashboard says it had disclosed 5,087 vulnerabilities across 680 open-source projects as of November 2, 2026. The dashboard reports that, to the company's knowledge, 629 had been patched; it also notes that a released patch does not guarantee widespread installation.

Maintainers of critical open-source projects can apply to OSS Scanner, a no-cost service that periodically scans their code and sends findings directly. Eligible maintainers and contributors can apply for six months of Claude Ultra at no cost through the Claude for Open Source Program. Security professionals, including open-source maintainers, can also apply to the Cyber Verification Program, which tailors Claude's safeguards for legitimate defensive work.

Sources

  1. Anthropic
  2. Anthropic
  3. Anthropic

Companies in this story

Related stories

Explore

More articles