This is step 3 in the Cognite Flows certification pipeline, sitting between code review and external app submission. It automates most of the tedious work in the official 10-question design quality assessment by running grep, lint, and build probes to detect hard-coded colors, missing ARIA labels, non-button clickables, and Aura token violations, then proposes a draft 1-5 score for each question before asking you to confirm or override. You still have to manually walk through your tasks in the running app since static analysis can't catch whether users get lost navigating screens. Targets a 3.8+ average to be launch-ready and writes a prioritized fix list to reviews/design-review. Solid time-saver if you're shipping Flows apps and tired of manually grading the same rubric.
npx -y skills add cognitedata/builder-skills --skill flows-design-review --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
github/awesome-copilot
microsoft/win-dev-skills
github/awesome-copilot