Context Compressor

Context compression for AI-assisted analysis is achieved through summary, outline, diff-only, and reference modes. These strategies maintain rich understanding while minimizing token usage by tracking file structure without loading entire files.

Summary mode provides a concise summary of files after full understanding, reducing the need for re-reading. This approach ensures minimal token usage while maintaining rich understanding for AI-assisted documentation review.

Diff-only mode focuses on changes in files, aiding in quick updates understanding. This context compression technique allows you to analyze code changes without loading the entire file, conserving tokens.

Progressive disclosure helps with information compression by allowing users to start with an outline and progressively load more detail as needed. This optimizes communication in contexts where token efficiency is critical for AI-assisted documentation review.

Reference mode tracks files without loading them, conserving tokens during file summarization. You should use this context compression strategy when you need to maintain awareness of file structure without processing the entire content.

Outline mode offers a high-level structure of files without delving into implementation details. This context compression strategy provides the structural overview needed for AI-assisted analysis while maximizing token efficiency.