emergency-response

What This Skill Does

The emergency-response skill is an essential framework designed to empower your OpenClaw agent to manage customer complaints, service disputes, and urgent operational issues with professional composure. It provides a structured hierarchy of incident levels (Level 1 to Level 4) that dictate how the agent should handle incoming stress, ranging from simple clarification requests to high-stakes legal risks. The skill incorporates proven communication templates, standardized conflict resolution workflows, and clear guidelines for when to escalate issues to a human supervisor. By formalizing the response process, it minimizes the risk of human-agent friction and ensures consistent, policy-compliant communication across all support channels.

Installation

To integrate this skill into your agent, run the following command in your terminal:

clawhub install openclaw/skills/skills/sa9saq/emergency-response

Use Cases

This skill is perfect for customer-facing AI agents operating on platforms like Coconala, e-commerce stores, or SaaS support portals. It is specifically useful for:

• Handling refund requests and ensuring they meet internal criteria.
• Managing communication during unexpected delivery delays.
• De-escalating tense customer interactions by following the structured 'Emotional Acknowledgement, Fact Confirmation, Response Declaration' workflow.
• Maintaining an audit trail of issues for future process optimization.

Example Prompts

1. "A client is demanding a refund because they feel the delivered project didn't meet their expectations. Please assess the situation based on the emergency-response guidelines and draft a professional, empathetic response."
2. "I am running behind schedule on project ID #9982. Please draft an apology message for the client, explaining the delay and providing a new estimated delivery date."
3. "The customer is threatening legal action over a misunderstanding of our terms. Trigger the Level 4 escalation protocol and draft the notification for my human supervisor."

Tips & Limitations

• Tip: Always encourage the agent to log every interaction in the post-processing phase. Consistent data collection is vital for long-term service improvements.
• Tip: Customize the 'Refund Decision Criteria' within the skill configuration to match your business's unique legal and operational requirements.
• Limitation: While the AI is trained on conflict resolution, it cannot physically perform financial transactions on platforms. It will always trigger an alert for the human to finalize the actual refund.
• Warning: Never allow the agent to issue refunds or negotiate contract terms independently; the skill is designed to support the decision-making process, not bypass human oversight.