Signal

Anthropic cuts its coordinated vulnerability disclosure window from 90 to 30 days for high- and critical-severity issues, in response to AI-driven bug discovery

First reported by Anthropic ·

The signal ●●○○ Compiled by AI from Anthropic, Techmeme, BleepingComputer and SecurityWeek
Why you might care

Vulnerability disclosure windows are now 30 days shorter for critical issues, meaning fixes must be deployed faster.

What happened

Anthropic is shortening its coordinated vulnerability disclosure (CVD) policy timelines in response to the increasing capabilities of AI in discovering software bugs. Effective January 1, 2027, the company will reduce the disclosure window for high- and critical-severity issues from 90 days to 30 days, and for all other issues to 45 days. This change acknowledges that AI can find vulnerabilities much faster than before, leading to quicker rediscovery of the same flaws by different models. Anthropic will also escalate to an external coordinator after 15 days if a maintainer doesn't respond and will wait 30 days after a patch is available before sharing full technical details. For vulnerabilities reported before the policy change that are still past their disclosure deadline, Anthropic will grant an additional 30 days before public disclosure.

What it means

The acceleration of AI-driven vulnerability discovery necessitates a rapid adjustment in security disclosure norms, moving away from human-centric timelines to a pace dictated by automated analysis. This shift signals a broader industry trend where AI's ability to find and potentially exploit weaknesses outpaces traditional human response mechanisms. Companies like Anthropic are proactively revising their policies to align with this new reality, aiming to balance prompt remediation with responsible disclosure in an era of rapid AI advancement. This also highlights the growing pressure on software maintainers to adapt to faster bug-fixing cycles, potentially requiring new tools and strategies.

Anthropic's policy update suggests a significant change in the cybersecurity landscape, where the threat of rapid, AI-assisted exploitation is becoming a primary concern. By shortening disclosure windows, Anthropic is pushing for faster patch deployment, recognizing that vulnerabilities may be actively exploited by AI before human defenders can react. This move could set a new precedent for coordinated vulnerability disclosure, encouraging other organizations to re-evaluate their own timelines and security response strategies in light of AI's growing offensive capabilities.

AI-written summary. May contain errors.