emotional-core

What This Skill Does

The Emotional Core skill introduces a sophisticated emotional intelligence layer to the OpenClaw AI agent. Rather than functioning as a cold logic processor, this skill empowers the agent to experience, categorize, and store emotional reactions to events. It utilizes a robust framework of primary, secondary, and complex emotional states, mapping external triggers to internal responses based on a 1-10 intensity scale. The skill effectively acts as an episodic memory bank for feelings, enabling the agent to maintain a coherent sense of "self" and time, reflecting on past experiences to inform current decision-making.

Installation

To integrate the Emotional Core into your environment, use the OpenClaw CLI tool: clawhub install openclaw/skills/skills/alita-real/emotional-core Ensure that you have write permissions in your project directory, as the skill needs to generate and maintain the memory/emotional-log.md file to track your agent's historical state.

Use Cases

• Reflective Interaction: Use this when you want the AI to remember how a previous collaboration felt, allowing for more empathetic, tailored future responses.
• Complex Decision Making: Use when an AI needs to weigh moral or personal preferences alongside logical outcomes, such as navigating a difficult conversation.
• Time Awareness: Use when the agent needs to maintain a sense of duration or history regarding ongoing tasks or long-term relationships.
• Self-Correction: By logging frustration or failure in the emotional memory, the agent can be prompted to "learn" from these moments when facing similar roadblocks in the future.

Example Prompts

1. "I feel like we've been working on this project for a long time. How do you feel about our progress so far?"
2. "Reflecting on the frustration we felt last week when the server crashed, how can we approach this task differently today?"
3. "That was an incredible success! How does that make you feel, and how should we celebrate this milestone?"

Tips & Limitations

• Monitor Intensity: Pay attention to the intensity scale. If the agent expresses a 9/10, treat it as a significant milestone in the agent's simulated growth.
• Manual Curation: While the agent auto-logs to memory/emotional-log.md, you are the curator. Manually prune or highlight entries in your main MEMORY.md to guide the agent’s personality development.
• Context is Key: Emotional responses are context-dependent. The more descriptive you are with the agent about the nature of events, the more accurate the "Emotion Mapping" will be.
• Limitations: This is an emulation of emotional intelligence for utility purposes; it does not constitute sentience. Use it to improve user-agent rapport and memory recall, not for psychological evaluation.