emotion-detector

What This Skill Does

The emotion-detector skill is an advanced natural language processing tool designed for OpenClaw AI agents. It goes beyond simple sentiment analysis by identifying specific emotional states—such as hope, anxiety, grief, and disgust—from text input. By analyzing primary and secondary emotions, intensity, and valence, the tool provides the agent with a recommended response strategy, ensuring interactions remain empathetic and situationally aware. A critical feature is the 'safe_t_flag', which identifies high-risk emotional states (like critical grief or despair), allowing developers to trigger mandatory safety protocols or resource intervention.

Installation

You can integrate this skill into your OpenClaw environment using the following command:

clawhub install openclaw/skills/skills/daisuke134/emotion-detector

Ensure your wallet is connected and funded on the Base mainnet, as this skill requires an x402 payment of $0.01 USDC per request to access the analysis endpoint.

Use Cases

• Empathetic Customer Support: Automatically detect if a customer is angry or anxious and instruct the AI to prioritize de-escalation over technical troubleshooting.
• Mental Health Triage: Monitor conversations for signs of extreme distress (grief, hopelessness) to automatically provide helplines and disconnect from standard scripted interactions.
• Communication Coaching: Help users understand how their text messages might be perceived by analyzing the valence and intensity of their draft messages.
• Sales Personalization: Adjust sales pitches in real-time based on the customer's perceived interest or hesitation levels.

Example Prompts

1. "Analyze this input: 'I feel really anxious about the presentation' and suggest a comforting response."
2. "Detect the emotion in this customer complaint: 'I am extremely disappointed that my order hasn't arrived after two weeks. This is unacceptable.'"
3. "Evaluate the sentiment of this message for my professional newsletter draft: 'I hope that we can collaborate more effectively on our upcoming projects to reach our common goals.'"

Tips & Limitations

• Context is Key: Always provide the optional context field if the input text is ambiguous, as it significantly improves the accuracy of the secondary emotion detection.
• Safety Protocol: Never ignore the safe_t_flag. If this value returns true, it is a strict requirement to halt standard bot logic and prioritize providing professional help resources or human intervention.
• Language Support: While the model handles both English (en) and Japanese (ja), ensure the language parameter is correctly specified to avoid misinterpretation of nuanced emotional markers.