Sandcastle
A project under development by Matt Pocock. Its specific functionality is not described in detail in the source material; it is used to empirically test code review Skills.
Timeline (last 90 days)
Well-known developer Matt Pocock put Cursor's "hardcore-grade code quality review" Skill to the test, running the review on his own real project, Sandcastle.
Matt Pocock used the Skill to review the five PRs most recently merged into main and got seven findings, five of which were solid.
The review found that the init service had ballooned past 1000 lines, mixing multiple responsibilities; the Skill gave a good splitting plan and also wanted to add a small generic function to delete about 20 lines of repeated boilerplate at once.
The review found that special-case checks for issue trackers were scattered across three layers, and the Skill suggested using code judo to fold the differences of custom trackers into the types; Matt Pocock rated it "both correct."
The review found inconsistent template args types, and the Skill suggested changing to a discriminated union or separate fields, but Matt Pocock considered this a false positive.
All Facts (5)
The review found that the init service had ballooned past 1000 lines, mixing multiple responsibilities; the Skill gave a good splitting plan and also wanted to add a small generic function to delete about 20 lines of repeated boilerplate at once.
50%UnverifiedThe review found that special-case checks for issue trackers were scattered across three layers, and the Skill suggested using code judo to fold the differences of custom trackers into the types; Matt Pocock rated it "both correct."
50%UnverifiedThe review found inconsistent template args types, and the Skill suggested changing to a discriminated union or separate fields, but Matt Pocock considered this a false positive.
50%UnverifiedWell-known developer Matt Pocock put Cursor's "hardcore-grade code quality review" Skill to the test, running the review on his own real project, Sandcastle.
50%UnverifiedMatt Pocock used the Skill to review the five PRs most recently merged into main and got seven findings, five of which were solid.
50%