memory-merger

The memory-merger skill is designed to streamline the process of consolidating mature lessons from a domain-specific memory file into its corresponding instruction file. By doing so, it ensures that valuable knowledge is preserved while reducing redundancy and maintaining clarity. This skill is particularly useful in environments where instructions and best practices are continuously evolving, such as in software development, project management, or AI prompt engineering. It automates the tedious process of reviewing and merging information manually, making knowledge management more efficient and reliable.

The skill provides several key features to facilitate this process. Users can specify the scope of memory merging, choosing between global (applied across all projects) and workspace (applied only to the current project) instructions. The workflow guides users through parsing input, reading relevant files, analyzing memory content, proposing mergers, defining a quality bar, and iteratively refining the merged instructions until they meet high standards for completeness, clarity, and minimal redundancy. It also ensures proper file updates, including frontmatter handling and merging of `applyTo` patterns, while cleaning up merged sections from memory files.

Memory-merger is ideal for developers, technical writers, and knowledge managers who need to maintain consistent and high-quality instruction files across multiple projects or teams. It is particularly beneficial in collaborative environments where lessons learned and best practices need to be integrated systematically. By using this skill, users can save time, reduce human error, and ensure that all relevant knowledge is captured and easily accessible, ultimately supporting better decision-making and more efficient workflows.