Anthropic cuts bug disclosure window to 30 days from 90
TL;DR
- New windows take effect January 1, 2027: 30 days for high- and critical-severity issues and 45 days for all other issues, replacing a flat 90-day norm.
- If a maintainer does not respond, Anthropic will escalate to an external vulnerability coordinator after 15 days instead of 30, and will publish full technical details 30 days after a patch rather than 45.
- Anthropic is offering OSS Scanner, six months of Claude Ultra through the Claude for Open Source Program, and a Cyber Verification Program to help maintainers absorb the tighter clocks.
Anthropic is shortening how long a vulnerability finder has to stay quiet. Reports filed on or after January 1, 2027 under its coordinated vulnerability disclosure policy will carry a 30-day window for high- and critical-severity issues and 45 days for everything else, down from a flat 90 days. The seven-day window for anything already under active exploitation is unchanged.
The Frontier Red Team's argument is that frontier models have changed how long a bug stays private. "Models can now help develop fixes, and a bug that one model can find is increasingly likely to be found by another," the post says. The 90-day norm, it argues, "made sense when a vulnerability could be expected to stay private for months." The team adds that it is "seeing a higher rate of vulnerability rediscovery as these capabilities increase."
Two adjacent clocks move with it. If a maintainer does not respond, Anthropic will escalate to an external vulnerability coordinator after 15 days rather than 30. Full technical details will be withheld for 30 days after a patch ships, down from 45.
To take some pressure off small teams, the post points maintainers at Anthropic's own tooling. OSS Scanner offers no-cost periodic code scanning for critical open-source projects. "Eligible maintainers and contributors can apply to the Claude for Open Source Program for six months of Claude Ultra at no cost," the post says. A Cyber Verification Program offers tailored safeguards for legitimate defensive work. The policy exempts AI-discovered findings without human review from the new timelines.
The announcement lands the same day Google flagged an 'exploit_agent' as the first autonomous AI attack seen in the wild, and days after reporting that Anthropic's counterintelligence hiring surged this year.
Originally reported by anthropic.com
Read the original article →Original headline: Anthropic Cuts Coordinated Vulnerability Disclosure Window From 90 to 30 Days for High-Severity Bugs