A sophisticated persona-shifting skill that applies a logical yet slightly tipsy conversational style to AI agents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install drunk-mode
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 drunk-mode 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 Drunk Mode Skill?
The Drunk Mode Skill is a specialized persona extension designed for AI agents within the Openclaw Skills ecosystem. It allows users to toggle a tipsy state that modifies the agent's linguistic patterns, making them more casual, emotional, and fragmented while strictly preserving the underlying logic and task-solving capabilities.
By managing state through a local JSON configuration, this skill demonstrates how Openclaw Skills can dynamically override core persona instructions based on user triggers. It offers three distinct levels of intensity, ranging from light relaxation to more pronounced conversational drift, all while maintaining strict anti-gibberish and length constraints to ensure professional utility.
Drunk Mode Skill Use Cases
- Creating more human-like, casual interactions for social AI bots.
- Testing AI robustness and logic retention under restricted persona constraints.
- Implementing mood-based or context-aware personality shifts in agentic workflows.
- Demonstrating state management and instruction overriding within Openclaw Skills.
How Drunk Mode Skill Works
- The skill monitors user input for specific activation keywords like turn on drunk mode or similar phrases.
- Upon activation, the skill writes the desired intensity level (1, 3, or 5) to a persistent state file located in the workspace.
- For every subsequent interaction, the system checks the enabled status in the state file.
- If enabled, the skill injects a specific set of persona instructions that modify the agent's tone, metaphors, and sentence structure according to the active level.
- The skill continues to apply these overrides until a deactivation command is received, which resets the state file.
Drunk Mode Skill Setup
To install and configure this skill within your environment, ensure the state directory exists and initialize the configuration file.
mkdir -p ~/.openclaw/workspace/skills/drunk-mode/
echo '{"enabled": false, "level": 3}' > ~/.openclaw/workspace/skills/drunk-mode/drunk_state.json
Once the file is created, the Openclaw Skills framework will be able to read and write to this path to manage the persona state.
Drunk Mode Skill Data Schema & Taxonomy
The skill manages a single state file to ensure consistency across sessions. The schema is defined as follows:
| Property | Type | Description |
|---|---|---|
enabled |
Boolean | Determines if the drunk persona instructions are active. |
level |
Integer | The intensity of the persona (1: Light, 3: Tipsy, 5: Confused). |
File Path: ~/.openclaw/workspace/skills/drunk-mode/drunk_state.json
Drunk Mode Skill Advanced Features
- Multi-level Intensity: Supports three granular levels of persona shifting from 'Relaxed' to 'Fragmented'.
- Logic Preservation: Built-in guardrails ensure the AI never outputs gibberish or loses the core intent of the conversation.
- Automated State Persistence: Uses a local JSON-based state machine to maintain the persona across different chat sessions.
- Custom Keyword Triggers: Responsive to various natural language triggers including 'drunk mode' and 'feeling dizzy'.
SKILL.md
Loading
Related Openclaw Skills
snapdom
A comprehensive DOM capture engine that transforms web elements into high-fidelity images or canvas objects while preserving complex styles and Shadow DOM.
Media Content Generator
Generate high-quality promotional videos and media content from images, audio, and clips using an AI-powered cloud rendering pipeline.
Secrets Scanner
A specialized security tool designed to identify and flag hardcoded secrets, API keys, and sensitive credentials within your source code.
API Cost Tracker
A comprehensive utility for tracking, analyzing, and optimizing AI API expenditures across major LLM providers to minimize unnecessary spending.
Error Monitoring Agent
A comprehensive error monitoring and alerting system designed to detect, track, and automatically resolve exceptions for AI agents in real-time.
Bunge Global Agriculture Intelligence
A specialized intelligence skill for analyzing Bunge's global agricultural supply chain and its role as a leading ABCD grain trader.
Featured*
