关于
The review-claudemd skill automates the refinement of your AI context by auditing previous interactions stored in your project history. It extracts recent conversations, employs Sonnet subagents to identify recurring patterns or violated instructions, and suggests concrete updates for both local and global CLAUDE.md files. This ensures that Claude continuously learns from your feedback and project evolution, reducing repetitive corrections and improving response accuracy over time by formalizing implicit knowledge into explicit documentation.
