# UX Review Loop Human-readable ota output is product surface, not incidental debug text. When help text, rich/plain text output, or docs that promise command behavior change, maintainers should review both the automated premium snapshots and a small live-command path. ## Automated gate Run: ```bash ota run ux-review ``` This keeps the snapshot-backed premium CLI surfaces stable: - root help - repo `doctor` - repo `doctor --plain` - repo `detect` - repo `explain` at narrow terminal widths - repo `up` - repo `run` - repo `agents` - workspace `validate` - workspace `doctor` - workspace `explain` - workspace `up` - workspace `run` ## Live review path Review these commands in a real terminal before shipping UX-sensitive changes: ```bash ota --help ota doctor . ota --plain doctor . ota detect --dry-run . env COLUMNS=48 ota explain . ota up . ota agents . ota run install:from:source ota workspace doctor . ota workspace explain . ota workspace up . ``` ## Review questions The output should stay: - calm - exact - grouped when repetition does not help - honest in `Why:` and `Next:` - visually consistent across repo and workspace surfaces - truthful about what ota can write, merge, or only review ## Decision rule If a change makes the CLI more technically complete but less legible in the first session, the change is not ready yet.