epistemology-limits

Classify epistemological limits and reframe unknowable questions into decision-relevant bounds.

212|23|Updated May 23, 2026
One-click install
npx skills add https://github.com/human-avatar/skills-for-humanity --skill epistemology-limits
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: epistemology-limits
Source: https://github.com/human-avatar/skills-for-humanity/tree/main/skills/epistemology-limits
Command: npx skills add https://github.com/human-avatar/skills-for-humanity --skill epistemology-limits

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you stop wasting effort on questions that cannot be settled and instead reframe them into precise, decision-relevant versions you can actually answer.

Core Features & Use Cases

  • Limit type classification: Distinguishes fundamental, practical, and conceptual limits (including underdetermination, observer effects, destroyed evidence, and category errors).
  • What remains knowable: Identifies bounds, proxies, and related sub-questions that can still be established under the limit.
  • Actionable reframing: Produces a clear reframed question (e.g., ranges, worst-cases, correlations, or confidence levels) so you can proceed responsibly despite uncertainty.
  • Use cases: Best for open scientific/analytic questions, investigations that have stalled, and situations where you’re asking “can this even be known?” rather than “what should we do next?”

Quick Start

Use the epistemology-limits skill to diagnose whether your question is fundamentally, practically, or conceptually unknowable and to produce a reframed, answerable version of it.

Frequently Asked Questions about epistemology-limits

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I reframe an unknowable question into actionable bounds for decision making?

To reframe an unknowable question, classify its epistemological limit type, identify remaining knowables like proxies or bounds, and output a decision-relevant question using ranges or worst-cases to proceed responsibly under uncertainty.

What is the best way to handle missing or inaccessible evidence in a stalled investigation?

Handling missing evidence involves classifying the practical limit, extracting what can still be established from related sub-questions, and producing a reframed question that relies on correlations or confidence levels rather than the destroyed evidence.

How do I deal with underdetermination between models when evidence is insufficient?

Dealing with underdetermination requires classifying it as a conceptual limit, stating the bounds that remain knowable across all competing models, and reframing the query to focus on practical significance rather than absolute model selection.

Can I get actionable answers when an observer effect prevents direct measurement?

Yes, you can get actionable answers by classifying the measurement interference as a fundamental limit, isolating the proxies that remain valid, and reframing the question to target correlations or confidence levels instead of direct observation.

When should I not try to settle an open analytic question and instead reframe it?

You should stop trying to settle an open analytic question when it is fundamentally, practically, or conceptually unknowable, and reframe it into precise bounds to extract what remains knowable for your decision making.