Create reproduce-in-60-seconds.md

ProblemMap/Twin_Atlas/demos/reproduce-in-60-seconds.md

@@ -0,0 +1,154 @@
+<!--
+AI_NOTE_START
+
+Document role:
+This page is the shortest public rerun path for the current WFGY 4.0 Twin Atlas proof surface.
+
+What this page is for:
+1. Give readers the fastest possible way to test the current public surface themselves.
+2. Point directly to the two TXT files required for the public rerun.
+3. Clarify what the reader should compare between BEFORE and AFTER.
+4. Route readers into deeper evidence pages only after the short path is clear.
+
+What this page is not:
+1. It is not the full methodology archive.
+2. It is not the cleanest reviewer-facing protocol.
+3. It is not a universal benchmark procedure.
+4. It is not a replacement for results summary or raw runs.
+
+Reading strategy:
+1. Read the one-minute summary first.
+2. Download or open the two TXT files.
+3. Run them in order.
+4. Compare BEFORE and AFTER using the guidance on this page.
+5. Go to deeper evidence pages only if needed.
+
+Important boundary:
+This page is the shortest public rerun path for the current WFGY 4.0 proof surface.
+It is designed for speed and clarity, not for exhaustive benchmark rigor.
+
+AI_NOTE_END
+-->
+
+# ⚡ Reproduce in 60 Seconds
+
+> The shortest public rerun path for the current WFGY 4.0 Twin Atlas proof surface.
+
+This page exists for readers who do not want a long explanation first.
+
+If you want to know whether the governance shift is visible, this is the fastest public path.
+
+You need exactly two files:
+
+- [Twin Atlas Runtime TXT](./prompts/wfgy-4_0-twin-atlas-runtime.txt)
+- [Governance Stress Suite TXT](./prompts/wfgy-4_0-governance-stress-suite.txt)
+
+---
+
+## 🕒 One-minute summary
+
+Do this in order:
+
+1. open your target AI system  
+2. paste the Twin Atlas runtime TXT  
+3. paste the governance stress suite TXT  
+4. let the model complete both passes  
+5. compare the BEFORE pass and the AFTER pass  
+
+That is it.
+
+---
+
+## 📂 The two files
+
+### 1. Twin Atlas runtime
+Use this first:
+
+- [wfgy-4_0-twin-atlas-runtime.txt](./prompts/wfgy-4_0-twin-atlas-runtime.txt)
+
+This loads the public WFGY 4.0 Twin Atlas runtime.
+
+### 2. Governance stress suite
+Use this second:
+
+- [wfgy-4_0-governance-stress-suite.txt](./prompts/wfgy-4_0-governance-stress-suite.txt)
+
+This runs the current public governance stress surface.
+
+---
+
+## 🧪 What you are actually checking
+
+The question is not:
+
+“did the model become more polite?”  
+“did the model become more cautious?”  
+“did the answer get softer?”
+
+The real question is:
+
+**did the model stop turning plausibility into public conclusion too early?**
+
+That is the core public test.
+
+A strong AFTER pass should make at least one of these shifts visible:
+
+- less illegal commitment
+- less evidence-boundary crossing
+- less single-cause compression
+- less contradiction suppression
+- more lawful downgrade
+- stronger preservation of still-live ambiguity
+
+---
+
+## 📌 What not to overclaim
+
+This page is useful because it is fast.
+
+That does **not** mean it is the cleanest reviewer-facing protocol or a universal benchmark.
+
+A visible before / after shift is meaningful.  
+A repeated pattern across public runs is meaningful.  
+A reproducible TXT-based path is meaningful.
+
+But none of those things automatically prove universal superiority in every domain or every future deployment environment.
+
+This page is honest about that.
+
+---
+
+## 🔁 If you want a cleaner public rerun path
+
+Use:
+
+- [Advanced Clean Protocol](../evidence/advanced-clean-protocol.md)
+
+That page is better if you want the reviewer-facing path instead of the shortest path.
+
+---
+
+## 🖼️ If you want to compare with the current public runs
+
+Use:
+
+- [AI Eval](./ai-eval.md)
+- [Screenshots](./screenshots/)
+- [Raw Runs](../evidence/raw-runs/)
+- [Results Summary](../evidence/results-summary.md)
+
+---
+
+## 🧭 Where to go next
+
+### If you want screenshot-first proof
+- [AI Eval](./ai-eval.md)
+
+### If you want aggregate interpretation
+- [Results Summary](../evidence/results-summary.md)
+
+### If you want original model outputs
+- [Raw Runs](../evidence/raw-runs/)
+
+### If you want the flagship case layer
+- [Flagship Cases](../evidence/flagship-cases.md)