A reusable /optimize skill measures speed first, applies safe fixes on a separate branch, reruns the same tests, and reports before and after plus what it couldn't measure.
An app feels slow and you want fixes checked against the same tests.
On a small internal app it found a duplicate page load and repeated data reads; repeat-visit time fell 32–40% per the author.
Save an /optimize skill covering N+1 queries, repeated reads, connection pooling, sequential API calls, heavy JavaScript and blocking tasks.
Have it measure first.
Apply safe fixes on a separate branch.
Rerun the same tests and read the before-and-after report.
What's good: measure, fix on a branch, retest. The catch: results are from the author's own app.
Write-ups follow public posts and official docs. No likes, views, or dashboard numbers. Open the source to check the original. About