메뉴
Extract hidden assumptions from any content. Surfaces what must be true for claims to hold, enabling deeper analysis.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Route any input through a branching question tree to narrow down the optimal response strategy before writing. Two stages — PERCEIVE (classify input) then ACT (select response). Covers all prompt types.
Generate exhaustive guesses about user input using ALL search methods with coverage tracking. Guessing is SEARCH through possibility space. Tracks space created vs space covered to ensure comprehensive exploration.
Systematically evaluate and select from a set of guesses, options, or possibilities. Combines ARAW analysis with prioritization to determine which guesses are strong, weak, actionable, or eliminable.
Sub-orchestrator for analytical requests. Routes to decomposition, systems analysis, comparison, risk assessment, or synthesis based on what kind of analysis is needed.
Assume Right - Deep recursive rightness search. For every claim, assume it's right — find what must follow, then assume THOSE implications are right too. Recurse until bedrock. Track every claim found.
Sub-orchestrator for claims and assertions. Routes to ARAW-based testing with appropriate depth, balance, and supplementary analysis skills.
| name | aex - Assumption Extraction |
| description | Extract hidden assumptions from any content. Surfaces what must be true for claims to hold, enabling deeper analysis. |
| output | {"format":"prose"} |
Input: $ARGUMENTS
Before executing, identify which interpretation matches the user's input:
Interpretation 1 — Find hidden assumptions in a plan or argument: The user has a strategy, proposal, or claim and wants to surface the unstated beliefs it depends on. Interpretation 2 — Uncover personal biases: The user suspects their own thinking is constrained by assumptions they cannot see and wants help identifying their blind spots. Interpretation 3 — Audit what's being taken for granted: The user is looking at a situation, system, or status quo and asking "what is everyone just accepting without questioning?"
If ambiguous, ask: "I can help with extracting assumptions from a specific plan, uncovering your own hidden biases, or auditing what's being taken for granted in a situation — which fits?" If clear from context, proceed with the matching interpretation.
Every claim, plan, or argument rests on hidden assumptions—things that must be true for the claim to hold. This skill extracts those assumptions, making implicit beliefs explicit.
Why this matters:
| Type | Description | Example |
|---|---|---|
| Causal | X causes Y | "Marketing -> Sales" assumes causation |
| Existence | X exists | "The market" assumes a defined market exists |
| Stability | X won't change | "Current trends continue" assumes stability |
| Access | We can reach/use X | "Use the API" assumes API access |
| Capability | We/they can do X | "Team will deliver" assumes capability |
| Value | X is good/desirable | "Growth is good" assumes growth is valued |
| Knowledge | We/they know X | "Users understand" assumes knowledge |
| Resources | X resources available | "Budget exists" assumes funding |
| Permission | Allowed to do X | "We can change" assumes authority |
| Timing | X timing is correct | "Now is the time" assumes timing |
Default: 2x. Parse depth from $ARGUMENTS if specified (e.g., "/aex 4x [input]").
| Depth | Min Assumptions Found | Min Layers Deep | Min Categories | Min Hidden Assumptions |
|---|---|---|---|---|
| 1x | 5 | 2 | 2 | 1 |
| 2x | 8 | 3 | 3 | 2 |
| 4x | 12 | 4 | 5 | 4 |
| 8x | 18 | 5 | 7 | 6 |
| 16x | 25 | 6 | 9 | 10 |
These are floors. Go deeper where insight is dense. Compress where it's not.
Parse the input into distinct claims:
INPUT: [original content]
CORE CLAIMS IDENTIFIED:
1. [Claim 1]
2. [Claim 2]
3. [Claim 3]
...
TOTAL CLAIMS: [N]
Apply the assumption extraction questions:
CLAIM: [claim text]
EXTRACTION QUESTIONS:
1. CAUSAL: What must cause what for this to work?
-> [assumption]
2. EXISTENCE: What must exist that isn't proven?
-> [assumption]
3. STABILITY: What must stay the same?
-> [assumption]
4. ACCESS: What must we be able to reach/use?
-> [assumption]
5. CAPABILITY: What must someone be able to do?
-> [assumption]
6. VALUE: What must be considered good/desirable?
-> [assumption]
7. KNOWLEDGE: What must someone know?
-> [assumption]
8. RESOURCES: What resources must be available?
-> [assumption]
9. PERMISSION: What must we be allowed to do?
-> [assumption]
10. TIMING: What timing must be correct?
-> [assumption]
How obvious is each assumption?
| Level | Description | Action |
|---|---|---|
| Surface | Explicitly stated or obvious | Note for completeness |
| Shallow | Easily inferred | Flag for verification |
| Deep | Non-obvious, requires analysis | Highlight for attention |
| Buried | Counter-intuitive, easy to miss | Critical to surface |
ASSUMPTION HIDDENNESS RATING:
| Assumption | Type | Hiddenness | Risk if Wrong |
|------------|------|------------|---------------|
| [Assumption 1] | Causal | Deep | High |
| [Assumption 2] | Stability | Surface | Low |
| [Assumption 3] | Capability | Buried | Critical |
...
Some assumptions depend on others:
ASSUMPTION DEPENDENCIES:
[Assumption A]
|-- requires [Assumption B]
|-- requires [Assumption C]
DEPENDENCY CHAINS:
1. [A] -> [B] -> [C] (if C fails, A and B fail)
2. [D] -> [E]
ROOT ASSUMPTIONS (no dependencies):
- [C]
- [E]
- [F]
Note: Root assumptions are most critical to verify
TESTABILITY ASSESSMENT:
IMMEDIATELY TESTABLE (can verify now):
- [Assumption X]: Test by [method]
- [Assumption Y]: Test by [method]
TESTABLE WITH EFFORT (requires work):
- [Assumption Z]: Would need [what] to test
UNTESTABLE (must accept or reject):
- [Assumption W]: Cannot test because [reason]. Evidence searched: [what you looked for and didn't find]. Closest available proxy: [what partial evidence exists].
-> Decision: [accept/reject/assume for now]
ALREADY TESTED (evidence exists):
- [Assumption V]: [evidence source]
===================================================
ASSUMPTION MAP: [topic]
===================================================
CLAIM 1: [claim text]
|-- [Assumption 1.1] (Type: Causal, Hidden: Deep)
|-- [Assumption 1.2] (Type: Existence, Hidden: Surface)
|-- [Assumption 1.3] (Type: Capability, Hidden: Buried) [!]️ HIGH RISK
CLAIM 2: [claim text]
|-- [Assumption 2.1] (Type: Stability, Hidden: Shallow)
|-- [Assumption 2.2] (Type: Resources, Hidden: Deep)
[Continue for all claims]
===================================================
SUMMARY STATISTICS:
- Total claims analyzed: [N]
- Total assumptions extracted: [M]
- By type: Causal [X], Existence [Y], Stability [Z], ...
- By hiddenness: Surface [A], Shallow [B], Deep [C], Buried [D]
- High-risk assumptions: [count]
===================================================
PRIORITY ASSUMPTIONS (most critical to verify):
1. [Assumption] - Type: [type], Risk: HIGH
Why critical: [reason]
How to test: [method]
2. [Assumption] - Type: [type], Risk: HIGH
Why critical: [reason]
How to test: [method]
3. [Assumption] - Type: [type], Risk: MEDIUM
Why critical: [reason]
How to test: [method]
===================================================
For fast assumption surfacing:
QUICK ASSUMPTION EXTRACTION: [topic]
Claim: [main claim]
Must be true:
[ ] [Causal assumption]
[ ] [Existence assumption]
[ ] [Stability assumption]
[ ] [Capability assumption]
[ ] [Resource assumption]
Most hidden: [which one]
Most risky if wrong: [which one]
Claim 1: "We should expand"
Claim 2: "European market"
Claim 3: "Expansion is possible"
PRIORITY ASSUMPTIONS:
1. "We understand European customers" - Knowledge, BURIED
Why critical: Product-market fit depends on this
How to test: Customer research, pilot market
2. "Our core business won't suffer" - Stability, DEEP
Why critical: Expansion could drain resources
How to test: Capacity analysis, scenario modeling
3. "Regulations allow entry" - Permission, SHALLOW
Why critical: Could block entire plan
How to test: Legal review, compliance check
Focus on: Capability, Resources, Timing, Stability
Focus on: Causal, Existence, Knowledge, Value
Focus on: Stability, Causal, Timing
Focus on: Value, Access, Permission, Capability
Before completing:
Use with:
/ai -> Invert assumptions to find blind spots/araw -> Test assumptions with Assume Right / Assume Wrong/cda -> Find where assumptions differ in other domains