Removing a cooldown lock that was never true to the source
The 7-day lock on active projects was idea-incubation logic pasted onto the wrong pipeline stage. Rotation already did the useful part of the job.
The 7-day lock on active projects was idea-incubation logic pasted onto the wrong pipeline stage. Rotation already did the useful part of the job.
Vaulting now really moves your files, and a "Keep Vaulted" no-op that wasn't - re-vaulting would have reset the exact timestamp the daily review sorts on.
Does advancing a stage count as a touch, what happens at the last stage, and a cooldown-vs-rotation tangent that turned out to be one question, not two.
axe-core over the existing component-test harness, scoped to actual WCAG rules - and an honest accounting of what a jsdom gate structurally cannot catch.
macOS has no official Tauri WebDriver. A community plugin fills the gap, but Cargo's feature unification nearly shipped its test server into release builds.
A five-finding cleanup batch where the smallest item mattered most: a regression test that failed because the bug it targeted is structurally unreachable.
An archived YAML crate, nine minor versions of rusqlite drift, and a duplicate thiserror major. The work was the reading, not the writing.
Dependency scanning without a new CI job, grouped update PRs, and the virtual-workspace detail that makes the cargo entry point at the wrong directory.
The UI arc: a wizard shell built before its steps existed, a capture flow torn out for being wrong, and the gap between green checkmarks and the truth.
The folder layout that outlives the app: zones versus stages, a config that refuses to overwrite itself, a database printout, and adopt-in-place import.
Page 3 of 4