Roman Gushchin has announced the
existence of an LLM-driven patch-review system named Sashiko. It automatically creates reviews
for all patches sent to the linux-kernel mailing list (and some others).
In my measurement, Sashiko was able to find 53% of bugs based on a
completely unfiltered set of 1,000 recent upstream issues using
“Fixes:” tags (using Gemini 3.1 Pro). Some might say that 53% is
not that impressive, but 100% of these issues were missed by human
reviewers.
Sashiko is built on Chris Mason’s review prompts (covered here in October 2025), but the
implementation has evolved considerably.