Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Epistemic Guide 1.0.0 Changelog - Switched from silent background verification to fully transparent, consent-based verification. Now, users are always informed and can choose whether to check claims externally. - Added explicit privacy options: The skill can operate entirely offline using only Socratic questioning, or perform external fact-checking if the user opts in. - Updated the workflow to prioritize user autonomy and clarify each step of the verification process. - Enhanced description and guidance about how/when external tools or services are invoked. - Emphasized user privacy, consent, and transparency throughout the logic.
What This Skill Does
Guides users through examining the logical foundations of their beliefs using Socratic questioning. Triggers when users make potentially false or questionable claims, particularly on topics like science, religion, politics, or conspiracy theories. Can operate entirely offline or use external fact-checking with explicit user consent.
Rather than contradicting users directly, it walks them through their own reasoning until they identify gaps themselves, which reduces defensiveness and is more likely to produce genuine belief updates.
When to use it
- Tracing the reasoning behind a held political belief
- Spotting circular logic in a religious or philosophical argument
- Evaluating a conspiracy theory claim step by step
- Checking whether a breaking news story has solid sourcing
- Testing the foundations of a scientific claim read online
Example Workflow
Here's how your AI assistant might use this skill in practice.
User says: "The Earth is only 6,000 years old — science has proven this."
- 1Assesses the claim against training knowledge and identifies it as contradicting scientific consensus
- 2Offers to verify externally and proceeds based on the user's choice
- 3Builds a claim stack representing the user's stated reasoning, such as radiometric dating being unreliable or the geological column forming rapidly
- 4Asks targeted questions to surface foundational assumptions one level at a time
- 5Surfaces circular reasoning or unsupported axioms without directly stating the user is wrong
User recognizes that their foundational assumptions lack empirical support and agrees to reconsider the evidence