Compress natural language markdown memory files to reduce token usage.
npx skills add https://github.com/kos3nz/dotfiles --skill caveman-compress-kos3nz
Please help me install this Agent Skill. Skill: caveman-compress Source: https://github.com/kos3nz/dotfiles/tree/main/.config/agents/skills/caveman-compress Command: npx skills add https://github.com/kos3nz/dotfiles --skill caveman-compress-kos3nz
Compress natural language memory files to reduce per-session token usage while preserving content integrity.
Run a natural-language markdown file through the caveman-compress workflow to reduce token load.
High-intent search queries and answers about installing and using this skill.
To reduce token usage from markdown memory files, you can compress prose files like CLAUDE.md using a deterministic compression workflow that preserves code blocks, URLs, and headings while minimizing context tokens during interactions.
The best way to compress markdown context files without losing code blocks is using a caveman compression workflow that preserves code blocks, URLs, headings, and file paths exactly while reducing natural language token load.
Yes, you can restore your original markdown file after token compression because the workflow automatically preserves backups as <filename>.original.md before applying any compression changes to the active memory file.
Markdown file compression works on memory files up to 500KB, enforcing a strict size limit and refusing to process sensitive filenames to ensure safe compression within Claude interaction contexts.
If markdown compression breaks your file formatting, the workflow validates outputs and retries up to two times with targeted fixes to ensure safe, accurate compression and content integrity.
You should not use token compression for markdown memory files if your file exceeds 500KB, contains sensitive filenames, or relies heavily on natural language prose nuances that cannot withstand deterministic compression.
selfagencyCompress markdown memory files into shorter phrasing while preserving code blocks and URLs.
huangzidaCompress natural language memory files into a compact caveman format.
g-lokCompress Markdown memory files into caveman format with backup.
monlorCompress natural language in markdown files while preserving code blocks and URLs.
