I write TypeScript and Next.js for a living, and GDScript when I'm making a game instead. On everything below, I was the only engineer, the only designer, and the only QA, so the verification had to be built into the work instead of into a reviewer. So a lot of what I've actually gotten better at isn't a framework or a language, it's noticing when I'm wrong before someone else has to.
That shows up as a habit more than a skill: write down the traps that cost real time, measure instead of assume, and say what I checked and how, rather than what I believe should be true. The case studies are mostly about the times that habit was missing and what it cost, and the one time it caught something before it shipped.