Explain

What This Skill Does

The Explain skill is a sophisticated cognitive layer for your OpenClaw agent designed to bridge the communication gap between AI logic and human understanding. It functions as a dynamic feedback loop that tracks how you interact with information, allowing the agent to refine its output format, depth, and pedagogical style over time. Rather than providing a static, one-size-fits-all response, the Explain skill categorizes your interactions by domain—such as code, abstract concepts, or debugging—and monitors success metrics. It observes when you affirm understanding versus when you follow up with clarifying questions, eventually building a personalized memory profile in ~/explain/memory.md that dictates exactly how the agent should present data to you in the future.

Installation

To integrate the Explain skill into your environment, use the OpenClaw CLI tool. Run the following command in your terminal to fetch the latest version from the official repository:

clawhub install openclaw/skills/skills/ivangdavila/explain

Ensure that you have write permissions enabled for your home directory, as the skill needs to create and maintain the ~/explain/memory.md file to persist your preference patterns.

Use Cases

• Technical Learning: As you learn a new framework, the skill adapts to provide more code snippets and fewer definitions as you signal competence.
• Complex Problem Solving: When debugging, the agent tracks whether you prefer step-by-step diagnostic lists or high-level architectural summaries.
• Knowledge Management: Use the agent to explain complex topics like finance or physics, where the skill will learn whether you prefer heavy technical jargon or intuitive real-world analogies.
• Iterative Refinement: If you find yourself consistently asking for shorter responses, the skill will recognize this pattern and adjust its default verbosity accordingly.

Example Prompts

1. "Explain how the Kubernetes ingress controller works, but keep it high-level since I'm just getting started."
2. "I don't understand that technical explanation. Can you rephrase it using an analogy related to construction instead?"
3. "That was too much detail. In the future, please provide only the summary and ask if I need a deep dive."

Tips & Limitations

• Signal Consistency: The agent requires at least two consistent signals to form a 'pattern' before adjusting its behavior. Providing clear, immediate feedback like 'got it' or 'too complex' helps the skill learn faster.
• Manual Overrides: While the skill auto-updates your ~/explain/memory.md, you are encouraged to manually edit this file to lock in specific preferences that you know you will never want to change.
• Scope: This skill is strictly for human-facing explanations. It will not change how the agent writes internal system logs or performs background tasks.
• Default Behavior: Until a preference reaches the 'confirmed' status, the skill defaults to providing direct answers first, followed by necessary context, ensuring you aren't overwhelmed by unexpected format changes.