This transforms code reviews from nitpicking sessions into actual knowledge sharing. It walks you through a phased review process (context gathering, high-level design, line-by-line, summary) and teaches you to give feedback that's educational rather than judmental. The severity labels (blocking, nit, suggestion, praise) are genuinely useful for setting expectations. It includes language-specific gotchas for Python and TypeScript, like mutable default arguments and the dangers of using "any". The question-based approach ("What happens if items is empty?") beats telling people what's wrong. If your team treats PRs as gatekeeping rather than collaboration, this reframes the whole practice around maintainability and learning.
npx -y skills add galaxy-dawn/claude-scholar --skill code-review-excellence --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
cursor/plugins
github/awesome-copilot
alirezarezvani/claude-skills
microsoft/win-dev-skills