GStars
    JuliusBrussee

    JuliusBrussee/caveman

    #42 this week

    🪨 why use many token when few token do trick. Viral skill + proxy for coding agents that cuts 65% of tokens by talking like a caveman.

    ai
    llm
    anthropic
    caveman
    claude
    claude-code
    meme
    Go
    Apache-2.0
    118.9K stars
    6.9K forks
    256 GitHub watchers
    Updated Nov 11, 2026
    View on GitHub

    Build with Backblaze B2

    SDKs, agent skills, IDE extensions, and reference pipelines from Backblaze Labs. All open source.

    Explore Backblaze Labs

    Loading star history...

    Where it stands

    Rank, growth pace and release cadence relative to everything GStars tracks.

    Ranks #7 among AI repositories ranked this week, and #10 over the last 30 days; #42 of 163 repositories with measured growth this week (top 26%).

    This week's +1,676 stars is in line with its 30-day pace of about +1,792 per week.

    5 releases in the last 90 days, the latest 2 days ago, with 111 contributors on GitHub.

    Peaked at #4 in the weekly ranking the week of Jul 19, its best in the past year.

    Use Cases & Benefits

    • Compresses AI agent output by converting verbose replies into concise, token-efficient caveman-style language while preserving full technical accuracy.
    • Reduces output token usage by about 65%, significantly lowering API costs and increasing response speed without sacrificing content quality.
    • Use for minimizing token consumption in Claude Code or similar AI agents to save costs during technical Q&A or code explanation sessions.
    • Use for rewriting memory files or session context to reduce input token size and extend effective context window in multi-turn conversations.
    • Use for generating brief, clear commit messages, PR review comments, or status reports to streamline developer communication workflows.

    Should you depend on it?

    License terms, project hygiene and risks, from GitHub metadata and GStars' snapshots.

    License

    Apache 2.0

    Permissive: commercial use, modification and redistribution are allowed with attribution; includes an explicit patent grant from contributors.

    Community files

    • README: present
    • Contributing guide: present
    • Code of conduct: present

    Watch-outs

    No watch-outs from GStars' checks.

    General guidance from repository metadata, not legal advice. How we score

    From the README

    why use many token when few do trick

    A Claude Code skill/plugin (also Codex, Gemini, Cursor, Windsurf, Cline, Copilot, 30+ more) that makes agent talk like caveman — cuts ~75% of output tokens, keeps full technical accuracy. Brain still big. Mouth small.

    Same fix. 75% less word. Brain still big.

    Pick your level of grunt — lite (drop filler), full (default caveman), ultra (telegraphic), or wenyan (classical Chinese, even shorter). One command switch. Cost go down forever.

    Speak your tongue. Caveman keep your language. You write Portuguese, caveman grunt Portuguese. Spanish, French, same. Compress the style, not the language. Code, command, error string stay exact.

    "Novo ref de objeto cada render. Prop inline = novo ref = re-render. Envolva com useMemo."

    One line. Find every agent. Install for each.

    ~30 seconds. Needs Node ≥18. Skip agent you no have. Safe to re-run.

    Trigger: type /caveman or say "talk like caveman". Stop with "normal mode".

    One agent only, manual command, or any of 30+ other agents → INSTALL.md. Install break? Open agent, say "Read CLAUDE.md and INSTALL.md, install caveman for me." Agent fix own brain.

    Excerpt from the JuliusBrussee/caveman README on GitHub (Apache-2.0 license). © its authors.Read the full README on GitHub

    Daily snapshots

    Stars gained per day, and how forks and open issues moved, from GStars' daily snapshots

    Aug 14: +87Aug 20: +304Sep 3: +472Sep 4: +489Sep 5: +286Sep 6: +96Sep 9: +202Sep 10: +184Sep 11: +274Sep 12: +113Sep 17: +344Sep 18: +249Sep 19: +195Sep 20: +87Sep 21: +300Sep 22: +213Sep 23: +183Sep 24: +69Oct 2: +281Oct 3: +467Oct 4: +331Oct 5: +197Oct 6: +203Oct 7: +202Oct 8: +173Oct 9: +181Oct 10: +166Oct 11: +55Oct 16: +642Oct 17: +514Oct 18: +240Oct 21: +185Oct 22: +274Oct 23: +280Oct 26: +222Oct 27: +258Oct 28: +232Oct 31: +229Nov 1: +192Nov 5: +296Nov 6: +275Nov 9: +205Nov 10: +272Nov 11: +242

    +20,907 stars over 90 days across 56 snapshot days; best day +642 on Oct 16.

    Forks

    6.9K+1.3K

    Open issues

    79-400

    Compared with peers in AI

    Fastest-growing repositories GStars tracks in this cohort for the week of Nov 8, by 7-day star growth

    View category →
    JuliusBrussee/caveman compared with AI repositories by stars, 7-day star growth, health score and last release
    RepositoryStars+7dHealthLast release
    D4Vinci/Scrapling101.1K+6.3K756 days ago
    thedotmack/claude-mem115.8K+4.7K75today
    heygen-com/hyperframes82.3K+4.0K75today
    harry0703/MoneyPrinterTurbo142.0K+3.8K753 days ago
    github/spec-kit149.2K+2.6K75today
    JuliusBrussee/caveman118.9K+1.7K752 days ago
    Category median115.8K+4.0K75—
    Compare caveman with Scrapling, claude-mem, hyperframes →

    Discover Repositories

    Search across tracked repositories by name or description