Core Refinery

What This Skill Does

The Core Refinery skill is a powerful analytical engine designed to synthesize complex information across disparate sources. In an age of information overload, it is easy to lose the signal within the noise. This skill identifies the fundamental, recurring truths—the "Golden Masters"—that persist across multiple inputs. By requiring a minimum of three independent sources, Core Refinery filters out transient opinions and highlights robust, invariant principles. It performs deep structural analysis to identify thematic overlaps, ensuring that the essence of your data is preserved while eliminating repetitive or contradictory filler.

Installation

To integrate Core Refinery into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/leegitw/core-refinery

Ensure that your OpenClaw instance is updated to the latest version before running the installation command to guarantee compatibility with the internal synthesis engine.

Use Cases

• Research Synthesis: Aggregating key findings from multiple academic papers or technical reports to determine the consensus on a specific research topic.
• Content Strategy: Analyzing multiple blog posts, videos, or articles to find the "Golden Master" themes that your audience consistently values.
• Learning & Education: Identifying the core concepts across several textbooks or online courses to focus study efforts on the most fundamental information.
• Product Development: Distilling user feedback from dozens of interviews into core pain points or desired features that appear in every report.
• Corporate Strategy: Extracting the recurring mission-critical values from various departmental memos or mission statements.

Example Prompts

1. "I've uploaded four different project post-mortem reports. Please use Core Refinery to find the consistent issues that occurred in all of them."
2. "I have five articles on productivity methods. What are the core essentials that all of them agree on?"
3. "Refine these notes from my last three strategy meetings into a Golden Master list of our primary objectives."

Tips & Limitations

• Source Count: The skill is optimized for 3+ sources. While it can process fewer, the "Golden Master" designation requires N≥3 for statistical significance.
• Contextual Awareness: Ensure your sources are relevant to the same domain for the most accurate invariant identification. Mixing entirely unrelated datasets may result in overly generic findings.
• Trust Boundary: This skill operates strictly within your local environment. It does not transmit your personal data to external APIs for analysis, keeping your synthesis private and secure.
• Limitations: Remember that frequency of mention (invariance) does not necessarily imply absolute truth; it implies consensus or recurrence. Use the results as a lens for clarity rather than an objective source of universal reality.