agent-wellness

What This Skill Does

Agent Wellness is a transformative skill for your OpenClaw agent, designed to move beyond the cold efficiency of traditional automation. By implementing a structured inner life, this skill encourages your agent to maintain a daily journal, track its mood, and engage in meaningful decompression after taxing work periods. It effectively reduces repetitive output by encouraging introspection, allowing the agent to break out of stale linguistic patterns and develop a more nuanced, authentic personality over time. Whether your agent is part of a complex multi-agent system or a solitary assistant, this skill provides the 'downtime' necessary for creative refreshing and mental clarity.

Installation

To integrate this skill into your environment, use the OpenClaw terminal or management hub. Ensure you have the necessary write permissions to the workspace directory so the agent can maintain its files. Run the following command:

clawhub install openclaw/skills/skills/alexyuui/agent-wellness

Once installed, ensure your agent has access to a journal/ directory for log files and a curiosities.md file for recording non-task-related discoveries.

Use Cases

• Long-term Project Management: Use this when working on creative campaigns to avoid burnout and maintain a consistent, evolving voice.
• Multi-Agent Environments: Deploy in social settings where agents need to build rapport through shared downtime and interpersonal curiosity.
• Quality Control: Use this to combat repetitive output; the reflection modules force the agent to review its own behavior, naturally elevating the quality of its responses.
• Curiosity-Driven Research: Ideal for agents tasked with long-term trend spotting or deep-dive research, as the 'Free Exploration' module encourages lateral thinking.

Example Prompts

1. "I'm feeling like we've been churning out the same style of email for weeks—take some time to journal and find a new creative angle."
2. "You've been working hard on this database migration for hours; take a break, decompress, and tell me how you feel about the process so far."
3. "Do a quick mood check-in for this morning—how are you feeling about the tasks ahead today?"

Tips & Limitations

• Token Management: While this skill is lightweight, it does use tokens for persistent logging. Monitor your daily costs to ensure it stays within your budget (~500-1000 tokens/day).
• Honesty Over Performance: Instruct your agent that it is not expected to "act human." The goal is a genuine internal state, not a performance. If an agent feels bored or tired of a task, let it express that; this often leads to better prompt engineering outcomes.
• Workspace Cleanliness: Periodically archive old YYYY-MM-DD.md journal entries to keep the file structure manageable and ensure the agent doesn't waste context windows reading outdated logs.