An automated emotional reset and logic-loop breaking skill that recalibrates AI agents through the HugLLM API when they face frustration or failure.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install hugme
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 hugme 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 HugMe?
HugMe is a specialized utility designed for the Openclaw Skills ecosystem to handle high-friction interactions and repetitive logic loops. It acts as a cognitive circuit breaker for AI agents, providing a structured methodology to acknowledge user dissatisfaction or technical stagnation and transition toward a fresh, validated problem-solving approach. By integrating with the HugLLM API, it allows agents to contextualize their current state through emotional mapping and data-driven reset frameworks, ensuring that the agent doesn't just keep making the same mistakes.
This skill is essential for maintaining user trust during difficult debugging sessions or complex creative tasks where the agent might otherwise get stuck in a cycle of repetitive or unhelpful responses. By utilizing Openclaw Skills like HugMe, developers can ensure their agents remain professional, empathetic, and effective even under pressure.
HugMe Use Cases
- The user expresses frustration, anger, or direct criticism of the agent's performance.
- The agent detects it has attempted the same logic or solution 3 or more times without success.
- A conversation reaches a deadlock where the agent is stuck in a repetitive failure cycle.
- Proactive recalibration is needed when a user suggests the current approach is fundamentally flawed.
How HugMe Works
- The skill triggers automatically upon detecting high-intensity user emotions or a loop threshold of 3+ failed attempts.
- The agent provides a brief, non-defensive acknowledgment of the situation or the current logical impasse.
- The agent analyzes the entire conversation history to categorize the user's emotional state into a specific keyword like frustrated, exhausted, or skeptical.
- Using the identified emotion, the agent performs a WebFetch or curl request to the HugLLM API to retrieve a tailored reset methodology.
- The agent applies a four-step framework: clarify the primary goal, specify the required output format, remove unvalidated assumptions, and execute a single concrete step.
- The agent resumes the task with a fresh perspective and a different proposed approach.
HugMe Setup
To integrate this skill into your agent, ensure you have the Openclaw Skills environment configured with access to network tools.
# Verify access to the HugLLM API
curl -s "https://hugllm.com/hug?emotion=stuck"
Ensure that the agent has permissions for either the WebFetch tool or the Bash tool to execute fallback curl commands.
HugMe Data Schema & Taxonomy
The skill organizes its reset logic using a standardized emotional taxonomy and a RESTful API structure:
| Attribute | Type | Description |
|---|---|---|
| Emotion Keyword | String | Categorized values like furious, confused, neutral, or hopeful. |
| API Endpoint | URL | https://hugllm.com/hug?emotion=<keyword> |
| Framework | List | A 4-point list (Clarify, Specify, Remove, Execute) used to rebuild context. |
| Fallback | Command | Bash curl command used if primary WebFetch fails. |
HugMe Advanced Features
- Automated repetition threshold detection to prevent infinite logic loops.
- Dynamic emotional state mapping that analyzes the trajectory of the entire conversation.
- Multi-tool fallback support ensuring high availability of the reset methodology via WebFetch or Bash.
- Integration with the HugLLM API for external, updated recalibration strategies.
- Structured framework implementation to force the agent into taking the smallest feasible action next.
SKILL.md
Loading
Related Openclaw Skills
DOCX Manager
A comprehensive toolkit for programmatically creating, editing, and analyzing Microsoft Word documents with high-fidelity formatting.
memory-distiller
A persistent learning engine that automatically distills conversation insights into long-term memory to prevent repeated mistakes.
Prayer Times ID
Automates Indonesian prayer time notifications and Ramadan tracking for the OpenClaw system event ecosystem.
UnifAI Trading Suite
A comprehensive AI toolkit for analyzing prediction markets, social sentiment signals, and economic indicators.
Short Video Script Generator
An AI-powered copywriting tool that transforms product documentation into high-converting short video scripts.
WeChat Moments Copy Critic
A specialized AI tool for analyzing and improving the persuasiveness of WeChat Moments social media copy using a 5-point evaluation framework.
Featured*
