Caveman Mode is an ultra-compressed communication protocol that cuts LLM token consumption by up to 75% by discarding conversational filler and articles while retaining 100% technical accuracy.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install mp-caveman
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install mp-caveman using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Caveman Mode?
Caveman Mode is a highly specialized developer prompt modification designed to maximize performance, response speeds, and token efficiency during interaction with AI coding assistants. By stripping away articles, conversational filler, polite intros, and unnecessary hedging, the agent communicates in a highly direct, terse, and objective-oriented manner.
This behavior is highly optimized for Openclaw Skills and developer workflows where time and token budgets are critical. Crucially, the system preserves raw technical data, code blocks, and error traces in their exact, uncompromised form, guaranteeing that speed does not come at the expense of accuracy.
Caveman Mode Use Cases
- API Token Optimization: Lower your token consumption by up to 75% during high-volume coding interactions, directly shrinking API bills.
- High-Speed Coding Cycles: Skip repetitive pleasantries and introductory boilerplate to get direct technical answers instantly.
- Terminal and CLI Environments: Read shorter, highly compact outputs that don't flood your console or narrow IDE panels.
- Rapid Debugging: Identify and resolve configuration or logic flaws immediately using quick causal mapping syntax.
How Caveman Mode Works
- Triggering: The protocol is initialized via user commands such as "start caveman" or "caveman mode".
- Syntactic Stripping: The AI agent filters out articles (a/an/the), pleasantries, and hedging, converting descriptions to short fragments.
- Causal Mapping: Ideas are structured using arrows (X -> Y) and common abbreviations (e.g., DB, auth, config, req, res).
- Technical Preserving: The model isolates code snippets and system errors, preventing any compression from altering execution syntax.
- Persistent Guard: The state is maintained actively through multi-turn prompts unless explicitly disabled by normal-mode commands.
Caveman Mode Setup
To activate Caveman Mode within your agent runtime, trigger the command directly within your agent interface:
# Trigger Caveman Mode manually
openclaw run skill caveman
To persist this mode, append the instruction set to your system configuration file. If you need to exit this ultra-compressed format, supply either of the following commands:
# Revert to standard conversational prose
stop caveman
# Or
normal mode
Caveman Mode Data Schema & Taxonomy
The skill manages its behavior dynamically and enforces structured constraints on output formatting as detailed below:
| Operational Element | Standard Verbose Mode | Caveman Mode Rule |
|---|---|---|
| Conversational Fillers | "Sure, let me help you rewrite that function to avoid errors..." | Strip completely. Get straight to the rewritten function block. |
| Syntactic Structures | "The token expired because the check is less than, not less than or equal to." | "Token expired. Check use < not <=. Fix:" |
| Database & Config | "You should check your configuration setup in your environment variable database." | "Check DB config env." |
| Code Blocks | Pretty-print output with text explanations. | Unchanged exact markdown code block, no pre/post-text. |
Caveman Mode Advanced Features
- Auto-Clarity Override: Automatically drops Caveman Mode temporarily when displaying critical security warnings, performing irreversible actions (like database drops), or outlining complex multi-step commands, immediately resuming once clear.
- Filler Drift Prevention: Employs robust context tracking within Openclaw Skills integrations to keep the AI from slipping back into polite prose during long, multi-turn chat sessions.
- Technical Integrity Guarantee: Forces absolute exactness on file paths, syntax, variable naming conventions, and package names while aggressively modifying prose around them.
SKILL.md
Loading
Related Openclaw Skills
AI Video Factory
An automated short-video generation skill that transforms a single topic prompt into a complete, high-quality video ready for multi-platform distribution.
Code Review and Quality Agent Skill
A multi-dimensional code quality gate and security audit skill that validates correctness, architecture, performance, and AI agent integrity before merge or deployment.
diagnose
A disciplined, multi-phase diagnosis loop designed for tracing hard bugs and performance regressions through structured reproduction and feedback loops.
Prototype Skill
A dedicated AI agent skill for building quick, throwaway prototypes to validate architectural logic or UI designs before committing to production code.
AI Trading Agent
A multi-agent AI trading companion providing 7 quantitative strategies, Kelly criterion position management, and robust risk control frameworks.
Todo Momentum
A high-momentum productivity tool that transforms overwhelming todo lists and paralyzing procrastination into immediate, low-friction physical actions.
Financial Wealth Management
An AI-powered wealth management and financial advisory command suite designed to automate client reporting, meeting preparation, financial planning, portfolio rebalancing, and tax-loss harvesting.
Featured*
