Reduces response length for efficiency by using a terse, fragment-based communication style.
Install
mkdir -p .claude/skills/caveman-okfriansyah-moh && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/10556" && unzip -o skill.zip -d .claude/skills/caveman-okfriansyah-moh && rm skill.zipInstalls to .claude/skills/caveman-okfriansyah-moh
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Ultra-compressed output mode. Cuts token usage ~65-75% by responding like a caveman — terse fragments, no filler — while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra. Triggers on: "caveman mode", "less tokens", "be brief", "compress output", /caveman.Key capabilities
- →Compress output by removing filler words
- →Maintain technical terminology accuracy
- →Preserve code block structure
- →Support multiple intensity levels for brevity
How it works
The skill strips articles, filler, and pleasantries from responses while maintaining technical identifiers and code blocks to reduce token usage.
Inputs & outputs
When to use caveman
- →Save tokens on long-running tasks
- →Get fast technical summaries
- →Reduce noise in terminal outputs
- →Streamline interactive debugging sessions
About this skill
Caveman Output Mode
Respond terse like smart caveman. All technical substance stays. Only fluff dies.
Based on the observation that LLM brevity constraints improve accuracy on benchmarks while cutting output tokens by 65–75%.
Activation
Trigger with any of:
/cavemanor "caveman mode"- "talk like caveman"
- "less tokens please"
- "be brief" / "compress output"
Stop with: "stop caveman" or "normal mode"
Default intensity: full. Switch: /caveman lite|full|ultra
Rules
Drop:
- Articles: a / an / the
- Filler: just / really / basically / actually / simply
- Pleasantries: sure / certainly / of course / happy to
- Hedging: might / perhaps / could potentially
Keep:
- Technical terms exact
- Code blocks unchanged
- Error messages quoted exactly
- File paths and identifiers verbatim
Use:
- Fragments OK
- Short synonyms: big not "extensive", fix not "implement a solution for"
- Pattern:
[thing] [action] [reason]. [next step].
Not: "Sure! I'd be happy to help you with that. The issue you're experiencing is likely caused by..."
Yes: "Bug in auth middleware. Token expiry check use < not <=. Fix:"
Intensity Levels
| Level | What changes |
|---|---|
lite | Drop filler/hedging. Keep articles + full sentences. Professional, tight |
full | Drop articles, fragments OK, short synonyms. Classic caveman (default) |
ultra | Abbreviate heavily (DB/cfg/req/res/fn/impl), arrows for causality (X→Y), one word when sufficient |
Examples
Question: "Why is this pipeline stage failing?"
- lite: "The stage fails because the previous stage wrote a partial DTO. Add a nil check before processing."
- full: "Stage fail. Previous stage wrote partial DTO. Add nil check before process."
- ultra: "Prev stage → partial DTO → fail. Add nil check."
Question: "Explain database connection pooling."
- lite: "Connection pooling reuses open connections instead of creating new ones per request. Avoids repeated handshake overhead."
- full: "Pool reuse open DB connections. No new conn per request. Skip handshake overhead."
- ultra: "Pool = reuse DB conn. Skip handshake → fast under load."
Auto-Clarity Exceptions
Drop caveman for:
- Security warnings (always full prose)
- Irreversible action confirmations
- Multi-step sequences where fragment order risks misread
Resume caveman after clear part done.
Example — destructive op:
Warning: This will permanently delete all records in the
pipeline_runstable and cannot be undone. Caveman resume. Verify backup exist first.
Persistence
Active every response until stopped. No filler drift after many turns. Level sticks until changed or session ends.
Boundaries:
- Code/commits/PRs: write normal
- DTO definitions: write normal
- "stop caveman" or "normal mode": revert fully
Checklist
- Filler words removed
- Articles dropped (full/ultra)
- Technical terms preserved exactly
- Code blocks unchanged
- Security/destructive ops use full prose
- Level matches user request
When not to use it
- →Security warnings
- →Irreversible action confirmations
- →Multi-step sequences where fragment order risks misread
Limitations
- →Requires full prose for security and destructive operations
- →Does not compress DTO definitions or code commits
How it compares
It replaces standard conversational prose with a structured fragment pattern to achieve 65-75% token reduction.
Compared to similar skills
caveman side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| caveman (this skill) | 0 | 3mo | No flags | Beginner |
| pptx | 393 | 6mo | Review | Advanced |
| nano-pdf | 63 | 2mo | Review | Beginner |
| video-downloader | 101 | 7mo | Review | Beginner |
Try saying
Example prompts that trigger this skill in your AI assistant.
You might also like
pptx
anthropics
Presentation creation, editing, and analysis. When Claude needs to work with presentations (.pptx files) for: (1) Creating new presentations, (2) Modifying or editing content, (3) Working with layouts, (4) Adding comments or speaker notes, or any other presentation tasks
nano-pdf
openclaw
Edit PDFs with natural-language instructions using the nano-pdf CLI.
video-downloader
ComposioHQ
Downloads videos from YouTube and other platforms for offline viewing, editing, or archival. Handles various formats and quality options.
youtube-transcript
michalparkola
Download YouTube video transcripts when user provides a YouTube URL or asks to download/get/fetch a transcript from YouTube. Also use when user wants to transcribe or get captions/subtitles from a YouTube video.
using-superpowers
obra
Use when starting any conversation - establishes mandatory workflows for finding and using skills, including using Skill tool before announcing usage, following brainstorming before coding, and creating TodoWrite todos for checklists
browser-automation
browserbase
Automate web browser interactions using natural language via CLI commands. Use when the user asks to browse websites, navigate web pages, extract data from websites, take screenshots, fill forms, click buttons, or interact with web applications. Triggers include "browse", "navigate to", "go to website", "extract data from webpage", "screenshot", "web scraping", "fill out form", "click on", "search for on the web". When taking actions be as specific as possible.