Stats
Actions
Tags
Help us improve
Share bugs, ideas, or general feedback.
From skills-for-humanity
Analyzes the nature, structure, and limits of knowledge. Routes to the appropriate sub-skill for clarifying what kind of claim is being made, what would justify belief, or how certain one can be. Use when investigating epistemic questions.
npx claudepluginhub human-avatar/skills-for-humanityHow this skill is triggered — by the user, by Claude, or both
Slash command
/skills-for-humanity:s4h-epistemologyThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Applies philosophical analysis to the nature, structure, and limits of knowledge. Diagnoses what kind of epistemic work is needed and routes to the right tool.
Classifies claims by type of knowing (a priori vs a posteriori, propositional vs procedural vs acquaintance) to determine justification standards. Use when assessing evidence or reasoning.
Calibrates AI confidence to evidence, flagging uncertainty and limitations before presenting conclusions. Useful when accuracy matters or knowledge is partial.
Enforces thinking disciplines for rigorous collaborative reasoning: map territory first, name confidence, sit with fog, verify before proposing, genuine agreement/disagreement. Auto-loaded by /figure-out, /define, and similar skills.
Share bugs, ideas, or general feedback.
Applies philosophical analysis to the nature, structure, and limits of knowledge. Diagnoses what kind of epistemic work is needed and routes to the right tool.
Epistemology is the meta-layer: not "is this claim true?" but "what kind of truth could this claim even have? What would justify believing it? How certain can we actually be?" It's most useful when you've hit a wall with ordinary investigation — when more evidence-gathering won't settle things, when you're not sure what kind of knowing is in play, or when you want honest calibration rather than confident-sounding output.
| You need to... | Tool |
|---|---|
| Map what kind of knowing is in play | epistemology-knowledge-types |
| Work out what would actually justify believing something | epistemology-justification |
| Get honest calibration: what do I know vs. believe vs. assume vs. hope? | epistemology-epistemic-status |
| Find out what can't be known, and why — and what can be established | epistemology-limits |
After diagnosing which tool fits, use the AskUserQuestion tool to confirm direction. Construct the question dynamically to include your diagnosis:
Proceed based on their selection.
Map what kind of knowing is actually in play — before testing whether the claim is true.
Identify the claim → classify what kind of knowing it invokes (a priori vs. a posteriori; propositional vs. procedural vs. acquaintance; sourced from perception, inference, testimony, or intuition) → assess what can and can't be established by that type. Different kinds of knowing have different standards of justification and different failure modes.
Output: Claim classified by knowledge type, with implications for what evidence is relevant and what the claim can and can't establish.
What would actually justify believing this?
Identify the belief → ask what would need to be true for it to be justified → classify the justification structure (foundationalist, coherentist, or reliabilist) → identify the weakest link in the chain. A belief can feel well-supported while resting on an unjustified foundation — justification analysis makes the load-bearing structure visible.
Output: Justification map with the belief's foundational support structure, classification of the justification type, and the specific point where the chain is weakest.
Honest, rigorous calibration: what do you know vs. believe vs. assume vs. hope?
Inventory all claims in a domain → assign each an epistemic status (known, reasonably believed, assumed, hoped, unknown) → trace dependencies: which high-confidence claims rest on lower-confidence foundations? Draws from the rationalist tradition of explicit epistemic labeling to replace confident-sounding vagueness with precise, honest calibration.
Output: Structured epistemic status map across the domain, with dependency chains flagged where confident claims rest on shakier ground.
What can't be known here, and why — and what can be established within those limits?
Identify what you're trying to know → classify the type of limit if one exists (fundamental: Gödel-style, underdetermination, observer effects; practical: evidence unavailable, destroyed, or counterfactual; conceptual: the question may be malformed) → clarify what can be established within those limits → reframe the question into the answerable part. The point is not to conclude "we can't know" — it's to be precise about what kind of knowing is and isn't available.
Output: Limit classification, what remains establishable, and a reframed question targeting the knowable.
After delivering this output, use AskUserQuestion to offer the next move:
/s4h-investigation — Apply practical investigation methods to what epistemology clarified/s4h-probability-confidence-calibration — Calibrate your confidence given the epistemic analysis/s4h-logic-check — Validate reasoning built on the epistemic foundations