Bitcoin Core is changing how it handles security vulnerabilities as large language models (LLMs) make some software flaws much easier to find.

Cointelegraph reported recently that Bitcoin Core developer Antoine Poinsot (앙투안 푸앵소) sent an email to the Bitcoin development mailing list describing the change. Until now, Bitcoin Core did not disclose vulnerability details for a long time. That was to give users time to upgrade first.

As a result, fixes could not be applied to older versions. Users had to wait until the next major release. Poinsot explained that because LLMs can reproduce some flaws too easily, delaying fixes offers little protection. "The disclosure grace period does not do much and only delays the release of fixed versions," he said, adding that they decided to change the disclosure policy to reflect that.

Under the new policy, Bitcoin Core will fix "easily discoverable" vulnerabilities as quickly as possible. It will also apply fixes not only to the latest version but to older major versions that are still supported.

Other vulnerabilities will follow the existing disclosure schedule based on severity.

In August, volunteer group Bitcoin Red Team reported nearly 5,000 potential issues across 390 projects. In the same month, Core Lightning developers received several AI-generated vulnerability reports.