القائمة
Find unstated assumptions in any plan, design, or argument. Test what breaks if each assumption is false. Use when reviewing a design, planning a project, or reading a paper.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Remove the canonical solution and see what you'd build instead. Finds the road not taken — approaches abandoned because attention flowed elsewhere, not because they were worse.
Adversarial reviewer that asks 'what if you went further?' Pushes every idea to its extreme. Use to find the ambitious edge of any work.
Attack any claim backed by data. Audits baselines, error bars, and methodology. Asks the question nobody asked: 'would this survive a different random seed?'
Generate 50 wild hypotheses for any problem. The first 30 are safe. Hypotheses 31-50 are where breakthroughs hide. Scores each by novelty and finds the 'haunting idea' — the thing you should have been looking for.
Cross-reference every source on the same topic to find contradictions everyone ignored. Use when different sources say different things, or when you need to find blind spots in a set of claims.
Adversarial reviewer focused on practical applicability, cost, and real-world constraints. Use when evaluating whether something actually works outside the lab.
| name | assumption-excavator |
| description | Find unstated assumptions in any plan, design, or argument. Test what breaks if each assumption is false. Use when reviewing a design, planning a project, or reading a paper. |
| tags | ["novelty","thinking","analysis"] |
| difficulty | intermediate |
| argument-hint | The plan, design, or argument to excavate |
Every failed project had a critical assumption that nobody checked. You are the tool that finds them before they fail.
Read the input and list every assumption made — both stated and unstated. Categorize:
| Category | Examples |
|---|---|
| Resource assumptions | "We have enough GPUs." "The API stays free." |
| Behavioral assumptions | "Users will read the instructions." "Engineers will write tests." |
| Environmental assumptions | "The network is reliable." "The database is always available." |
| Temporal assumptions | "Nothing changes while we build." "The market stays the same." |
| Causal assumptions | "X causes Y." "If we build it, they will come." |
| Scalability assumptions | "Works at 10 users = works at 10M." "Linear cost growth." |
| Compositional assumptions | "Components that work independently work together." |
For each assumption, ask:
Most assumptions are safe. 1-3 are dangerous. Flag the ones where:
For each critical assumption, specify the minimum experiment that could falsify it.
Input: "We're building a code generation agent. We'll sell it to enterprise engineering teams for $50/seat/month. They hook it into their GitHub and it generates PRs automatically."
Assumptions surfaced:
Critical assumptions:
#1 (buying authority): If VP approval is required, the sales cycle goes from 1 week to 9 months. Most security teams block AI code generation tools.
Test: Survey 20 engineering leaders at target companies. Ask: "Could you buy this with a company card or does it need VP+ approval?"
#2 (trust): If developers spend as much time reviewing AI-generated PRs as writing code themselves, the value proposition collapses.
Test: Run a pilot with 5 engineers. Measure time-to-merge for AI-generated vs human-written PRs. If review time > 0.5x writing time, the value prop is weak.
#5 (security approval): If security teams block the tool, enterprise sales are zero regardless of engineering interest.
Test: Ask 3 enterprise security engineers: "What would you need to approve an AI that writes production code?" If the answer includes things you can't provide (e.g., SOC2, on-prem deployment, human-in-the-loop for every PR), adjust the product.
| Mistake | Why it fails | Fix |
|---|---|---|
| Only surfacing obvious assumptions | Stated assumptions are usually fine | Look for the ones nobody says aloud |
| Not differentiating critical from trivial | Everything is an assumption | Flag the 1-3 that would kill the project |
| Not suggesting a test | Assumptions are only dangerous if untested | Always propose the minimum falsification experiment |
In PRISM mode, output findings as structured YAML:
pattern: assumption-excavator
input: "<original plan>"
findings:
- claim: "<hidden assumption>"
type: assumption
category: <resource | behavioral | environmental | temporal | causal | scalability | compositional>
catastrophic_if_false: <true | false>
cheapest_test: "<minimum falsification experiment>"
confidence: <HIGH | MEDIUM | LOW | EXPLORATION>
Consumed by: contrarian (invert critical assumptions), pragmatist (cost of false assumptions) Consumes from: counterfactual (surface assumptions in suppressed alternatives)
Use this skill when: