Principle Comparator

What This Skill Does

The Principle Comparator is a specialized analytical tool for OpenClaw that enables agents to perform cross-source synthesis. Rather than simply summarizing two texts, this skill performs a rigorous comparison between two inputs—whether they are raw text or pre-processed extractions—to identify shared and divergent principles. By normalizing concepts into their core values, the agent can track whether an idea is a singular observation (N=1), a validated principle across two sources (N=2), or a potential invariant (N≥3). This process helps users strip away stylistic differences to reveal the underlying architecture of the ideas being analyzed.

Installation

To integrate this skill into your environment, use the OpenClaw command-line interface: clawhub install openclaw/skills/skills/leegitw/principle-comparator Ensure your agent has the necessary permissions to process text segments and perform semantic normalization using your configured LLM.

Use Cases

This tool is ideal for researchers, analysts, and anyone performing comparative studies. Use it when you need to:

• Validate a hypothesis: Check if a specific claim in one document is supported by another.
• Synthesize research: Extract the universal truths from two conflicting or complementary perspectives.
• Filter noise: Determine which principles consistently appear across diverse datasets.
• Evaluate consistency: See if your own documentation, notes, or project plans align with external inputs.

Example Prompts

• "I have two articles about remote work culture. Use the Principle Comparator to find the core values they both share and where they diverge significantly."
• "Compare these two extraction files from my previous meeting notes. What are the invariant principles that appear in both sessions?"
• "I'm validating a new company mission statement. Compare my draft against this industry best-practice document and highlight any conflicting principles."

Tips & Limitations

• Semantic Alignment: The skill relies on normalization. If you feel a match is incorrect, verify the underlying logic; the agent is looking for core value, not just keyword overlap.
• Source Quality: This skill works best when inputs have already passed through an extractor like pbe-extractor, as it significantly reduces the amount of fluff the model needs to ignore.
• Scope: It does not evaluate 'correctness.' It is a logic engine, not a moral arbiter. It will tell you that two people agree on a bad idea, but it will not tell you the idea is inherently wrong or right.