A doc-drift fix that wasn't as boring as it sounded
Three audit items that each turned into something: a half-fixed claim, a quietly dead password reset, and a staging email that would have linked to production.
Three audit items that each turned into something: a half-fixed claim, a quietly dead password reset, and a staging email that would have linked to production.
Two coverage gaps closed, and the more useful question a review pass asked afterward: not "does this pass" but "what would this test actually catch."
A one-word fix to the login return-path that a scanner and a review each caught being incomplete: HEAD requests slipping past, and a stale stored URL nobody cleared.
A stale task that would have undone the CI migration, a master-key fix that escalated from shrink-the-window to close-it, and a parser that silently eats plausible syntax.
The jobs dashboard returned a 500 instead of a login page: bare route helpers resolve against the engine, not the app. One line, plus its dormant twin.
A cleanup issue with a wrong rationale, a docs sweep that wasn't needed, and the orphaned credential a review caught.
Dead CI, live deploys: folding every job onto the homelab runner, deleting the hosted-runner compensation machinery, and the CVE backlog waiting behind the gate.
A one-time cleanup fixed 80 posts and promised nothing. The second fix is twelve lines that fail the build, with zero exceptions.
Medium's native integration was a museum exhibit. A browser bridge, an article whose body is a URL, and minus 1,603 lines.
A whole dark theme by redefining CSS variables under one class, and 45 minutes spent proving a five-line toggle innocent.
Page 3 of 6