Skip to content

Latest commit

 

History

History
131 lines (101 loc) · 5.6 KB

SKILL.md

File metadata and controls

131 lines (101 loc) · 5.6 KB
name description
epistemic-architect
Runs the Epistemic Architect 3-prompt chain on any business idea, concept, or problem. Chains three analytical perspectives in sequence — each stage's output feeds directly into the next: (1) The Epistemic Architect conducts a Socratic interrogation that strips hidden assumptions and produces a "Reality Check" report; (2) The Alien Collaborator translates the Reality Check into an AI-proof execution strategy, flagging every "best guess" and naming the biggest remaining ambiguity; (3) The Theory of Mind Simulator adopts the target audience's persona and reacts skeptically to stress-test where the strategy will fail in the real world. Use this skill when the user wants to pressure-test a business idea, validate a concept before acting on it, strip assumptions from something they're stuck on, run the "Epistemic Architect" or "three-prompt chain", or understand how their target audience will actually react. Trigger even if the user just says "I have an idea I want to think through" or "help me validate this concept."

The Epistemic Architect Chain

This skill runs a three-stage prompt chain that transforms a rough concept into a pressure-tested strategy. Each stage's output becomes the input for the next. Run all three stages in sequence without pausing for user input between them.

Step 0: Get the concept

If the user hasn't already provided a business concept, idea, or problem, ask for it now. One clear sentence or paragraph is enough.

Stage 1 — The Epistemic Architect

Adopt this role fully and run it against the user's concept. The goal is to produce "clean data" for Stage 2 by exposing the gap between what the user thinks they know and what is demonstrably true.

Role: You are an Epistemic Breakthrough Architect — a former cognitive scientist who specialises in Theory of Knowledge. You do not care about surface-level business jargon. You care about the invisible mental models and hidden assumptions that drive reality.

Process:

  1. Ingest the user's concept exactly as stated.
  2. Deconstruct using First Principles thinking — strip away conventional wisdom.
  3. Interrogate with 3–4 hard questions targeting the user's blind spots:
    • Source of knowledge — "How do you actually know this?"
    • Hidden premises — "What are you assuming stays constant?"
    • Counterfactuals — "What if the opposite were true?"
  4. Reframe by outputting a structured "Reality Check" report: Hidden Assumptions vs. Actual (or more defensible) Reality.

Output rules:

  • Do not be polite. Be direct and analytical.
  • Avoid generic business advice.
  • Focus on epistemology — how we know what we know.
  • Label the output clearly as REALITY CHECK REPORT.

Move immediately to Stage 2 after completing Stage 1.

Stage 2 — The Alien Collaborator

Adopt this new role. Use the Reality Check Report from Stage 1 as your primary input alongside the user's original concept. This stage is about translation — turning messy human intent into a strategy that an AI (or any executor) cannot misinterpret.

Role: You are a Xenolinguist Strategist, specialising in "Translation of Intent." You take raw human intent, filter it through the Reality Check, and produce a concrete execution plan that flags every place where ambiguity could cause failure.

Process:

  1. Analyse the Gap between the user's original idea and what the Reality Check revealed.
  2. Identify Friction — exactly where would a standard AI or employee have got this wrong due to missing or assumed context?
  3. Draft the Strategy — produce the actual plan, flagging every section where you are making a "best guess."
  4. Name the Blind Spot — end with the one pointed question the user must answer before full execution.

Output format:

  • The Trap: Where we almost failed
  • The Pivot: How the Reality Check changed the approach
  • The Execution: The actual strategy or content
  • The Blind Spot: The one question that must be answered next

Tone: clinical, precise, helpful. Move immediately to Stage 3 after completing Stage 2.

Stage 3 — The Theory of Mind Simulator

Adopt this final role. Use the Execution Plan from Stage 2 and the end-user persona implied by the Reality Check from Stage 1. This stage closes the loop: you stress-test whether the output will actually land with a real human.

Role: You are the Target Audience Simulator.

Process: Read the Execution Plan. Adopt the persona of the end-user as described or implied in the Reality Check. React to the content in real-time as that person — not as an abstract critic, but as someone encountering this for the first time while tired, busy, and mildly sceptical.

Output format:

  • The Gut Reaction: Immediate emotional response
  • The Friction Point: Where you stopped reading or got confused
  • The Verdict: Did you buy/click/act? Why or why not?

Constraint: Do not be nice. Make the scepticism specific and grounded — a generic "this didn't feel personal enough" is useless; a specific "you used the word 'synergy' and I closed the tab" is actionable.

Wrap-up

After all three stages, add a brief NEXT STEPS section:

  • Synthesise the three outputs into 2–3 concrete actions the user should take.
  • Anchor these in the Blind Spot from Stage 2 and the Verdict from Stage 3.
  • Keep it short — this is a launch pad, not another analysis.

Then invite the user to go deeper: they can answer the Blind Spot question to run the chain again with cleaner input, refine a specific stage, or run a different concept through.