🎯Skills7
An ultra-compressed communication skill for Claude Code that reduces token usage by approximately 75% while maintaining 100% technical accuracy, optimized for Japanese language patterns including keigo removal and filler elimination.
A sub-skill of genshijin that generates concise Conventional Commits messages with subjects under 50 characters focused on the 'why', part of the Japanese-optimized token compression system that reduces usage by approximately 75%.
Compresses natural language memory files (CLAUDE.md, todos, configs) into a terse "caveman" format to reduce input tokens, while fully preserving code blocks, URLs, file paths, and technical content. The original human-readable version is backed up before overwriting.
A quick-reference skill for the genshijin concise output mode in Claude Code, offering three verbosity levels (polite, normal, extreme) that progressively strip Japanese grammar down to keyword-only responses, plus sub-skills for compact commit messages and one-line PR reviews.
A sub-skill of genshijin that generates ultra-compressed one-line PR review comments, part of the Japanese-optimized token compression system that reduces token usage by approximately 75% while maintaining technical accuracy.
An ultra-compressed communication skill for Claude Code optimized for Japanese, reducing token usage by approximately 75% by eliminating honorifics, filler words, and verbose expressions while maintaining technical accuracy.
A super-compressed communication skill for Claude Code optimized for Japanese, reducing token usage by approximately 75% while maintaining full technical accuracy. Includes three intensity levels and sub-skills for commits, PR reviews, and memory compression.