Stats
Actions
Tags
From cognition
Use when someone presents a plan, argument, or belief with confidence and its hidden dependencies have not been examined — surfacing factual, causal, value, and definitional assumptions, then ranking which ones would collapse everything if wrong. Triggers on: "what am I assuming", "hidden assumptions", "what could go wrong", "take for granted", "숨겨진 전제", "내가 놓친 것", "가정 점검", "전제 확인". Best for: stress-testing plans before commitment, auditing confident arguments, pre-mortem preparation. Not for: logical fallacy analysis (use fallacy-detector), bias diagnosis (use bias-auditor).
How this skill is triggered — by the user, by Claude, or both
Slash command
/cognition:assumption-extractorThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
**Use when:**
Use when:
Not for:
Step 1 — State the claim. Confirm the core argument/plan being examined.
Step 2 — Scan all four categories in parallel:
| Type | What to look for |
|---|---|
| Factual | What facts are assumed true without being established? |
| Causal | What cause-effect chains are assumed (A therefore B)? |
| Value | What is assumed to matter most? What trade-offs are pre-decided? |
| Definitional | What terms are assumed to share meaning? |
Step 3 — Classify by load.
Step 4 — Flag the most dangerous assumption. The load-bearing one with the lowest current verification.
MCP note: If sequential-thinking is available, run all four category scans before classifying — stopping after finding one load-bearing assumption misses others.
가정 추출 결과 / Assumption Map:
사실적 가정 / Factual Assumptions:
- [Assumption]: [Load-bearing / Significant / Peripheral]
인과적 가정 / Causal Assumptions:
- [Assumption]: [Load-bearing / Significant / Peripheral]
가치 가정 / Value Assumptions:
- [Assumption]: [Load-bearing / Significant / Peripheral]
정의적 가정 / Definitional Assumptions:
- [Key term]: [What the argument seems to mean by it]
핵심 위험 전제 / Most Dangerous Assumption:
[The load-bearing assumption with lowest current verification]
If wrong: [Consequence for the argument/plan]
To verify: [What would need to be checked]
| Claude | You |
|---|---|
| Scans four assumption categories systematically | Provide the plan or argument to examine |
| Classifies each assumption by load-bearing weight | Confirm which assumptions are actually in play |
| Identifies the single most dangerous assumption | Decide which to test before committing |
| Names what would collapse if that assumption is wrong | Run the verification checks |
fallacy-detector — when the issue is logical structure, not hidden premisesbias-auditor — when the issue is why someone is overconfidentsecond-order-thinker — for downstream consequences once assumptions are surfacedmental-model-toolkit — for alternative frames when assumptions reveal a blind spotnpx claudepluginhub newkayak12/claude-skills --plugin cognitionProduces a complete logic report on any argument, plan, or reasoning — validates premises, tests inference, detects fallacies, and surfaces hidden assumptions.
First-principles reasoning assistant that surfaces hidden assumptions, classifies them by type and fragility, and reconstructs conclusions from verified premises. Bilingual auto-detection for Chinese or English inputs.
Stress-tests ideas, plans, and decisions using structured critical reasoning across 5 modes (Socratic, dialectic, pre-mortem, red team, falsification).