Using AI tooling for code review

I found this interesting: https://github.com/spaceraccoon/vulnerability-spoiler-alert

Reasons:

  1. It’s not easy to fix the vulnerabilities in open source projects like bitcoin core.

    https://xcancel.com/fanquake/status/1977396233237864530 https://xcancel.com/christine_dkim/status/1978121490336796945

  2. The results in this experiment look good enough based on some of the repositories it scans everyday.

Questions:

  1. How would this affect bitcoin core and other open source bitcoin projects?
  2. Has there ever been an incident in bitcoin core where a vulnerability was exploited before a release because it was easy to spot the fix in the commits?

Note: Such reviews can also be used to find bugs in the commits and not just the bug fix.

Archive