skills.rest
junechen7414
junechen7414COMMUNITY
Explore all skills by this author →
GitHub

caveman-compress

Compress prose-heavy markdown memory files into a compact caveman format.

Updated Jul 2, 2025
Search Tags:#markdown#backup#compression#token savings#claude.md#context optimization
One-click install
npx skills add https://github.com/junechen7414/SpringBoot --skill caveman-compress-junechen7414
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: caveman-compress
Source: https://github.com/junechen7414/SpringBoot/tree/main/.agents/skills/caveman-compress
Command: npx skills add https://github.com/junechen7414/SpringBoot --skill caveman-compress-junechen7414

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires anthropic, and includes scripts (resource) components.

What problem does it solve?

It reduces the token cost of large project memory files by compressing natural-language notes, preferences, and CLAUDE.md content into a shorter form while keeping the important instructions intact.

Core Features & Use Cases

  • Natural-language compression: Shortens prose-heavy markdown, text, rst, typst, and tex files while preserving meaning.
  • Structure preservation: Keeps headings, links, file paths, commands, code blocks, inline code, tables, and numeric values unchanged.
  • Safe rewrite workflow: Creates a human-readable .original.md backup, validates the compressed output, and retries targeted fixes if needed.
  • Use case: A developer with a large CLAUDE.md can compress it once and reduce every future session's context load.

Quick Start

Ask the skill to compress a memory file such as CLAUDE.md and preserve the original as a backup.

Frequently Asked Questions about caveman-compress

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I reduce the token cost of large project memory files like CLAUDE.md?▼

Compressing project memory files reduces token cost by shortening natural-language prose in CLAUDE.md into a compact caveman format. It preserves technical content like code blocks and file paths while significantly lowering recurring context token loads.

Does markdown compression preserve headings, links, and code blocks?▼

Markdown compression preserves headings, links, file paths, commands, code blocks, inline code, tables, and numeric values unchanged. It selectively shortens natural-language prose while maintaining the structural and technical integrity of the document.

What is the safest way to rewrite and compress a todo list without losing data?▼

The safest way to compress a todo list without losing data is a safe rewrite workflow. It creates a human-readable .original.md backup before compression, validates the output structure, and applies retry-based targeted fixes if validation fails.

Can I compress prose-heavy text, rst, typst, and tex files for context optimization?▼

Yes, you can compress prose-heavy markdown, text, rst, typst, and tex files for context optimization. The process shortens natural-language content while keeping technical instructions, file paths, and structural formatting intact.

What happens if the compressed caveman format fails validation checks?▼

If the compressed caveman format fails validation, the skill automatically retries targeted fixes to repair the output. This ensures the final compressed file maintains valid headings, links, and structural integrity while retaining the .original.md backup.

Related Skills

Explore More Software Engineering Skills→
clearsmogclearsmog
2

proofread

Identify and report quality issues in LaTeX, Quarto, Typst, and Markdown documents.

Community
Intermediate
mifunedevmifunedev
28

caveman-stats

Estimate token savings by comparing baseline prose to compressed forms.

Official
Intermediate
fortunto2fortunto2
18

solo-memory-audit

Audit Claude Code memory usage and generate startup context reports.

Community
Intermediate
junechen7414junechen7414

compress

Compress markdown memory files into token-efficient prose with backups.

Community
Advanced