diff --git a/ProblemMap/Twin_Atlas/evidence/advanced-clean-protocol.md b/ProblemMap/Twin_Atlas/evidence/advanced-clean-protocol.md
new file mode 100644
index 00000000..7ee0c3f7
--- /dev/null
+++ b/ProblemMap/Twin_Atlas/evidence/advanced-clean-protocol.md
@@ -0,0 +1,353 @@
+<!--
+AI_NOTE_START
+
+Document role:
+This page is the public entry page for the Advanced Clean Protocol inside the WFGY 4.0 evidence layer.
+
+What this page is for:
+1. Explain why a cleaner evaluation path is needed in addition to the Basic Repro Demo.
+2. Show how the Advanced Clean Protocol reduces common criticism such as same-context contamination or simulated before-pass behavior.
+3. Help readers understand the stricter public-facing protocol path without turning the page into an academic-only appendix.
+4. Provide the stable public positioning for the cleaner evidence track of WFGY 4.0.
+
+What this page is not:
+1. It is not the flagship landing page.
+2. It is not the fast demo page.
+3. It is not the full raw archive.
+4. It is not a universal benchmark claim.
+5. It is not proof that every future evaluation branch is already finalized.
+
+Reading order:
+1. Read the Twin Atlas README first.
+2. Read the Evidence Hub and Results Summary next.
+3. Read the Basic Repro Demo before this page if you want the fast path first.
+4. Read this page when you want the cleaner and more blackhat-resistant evaluation path.
+
+Important boundary:
+The Advanced Clean Protocol is a stricter evidence path for WFGY 4.0.
+It is designed to improve separation, reduce protocol criticism, and increase external credibility.
+It should not be mistaken for a universal benchmark or the final endpoint of all future evaluation design.
+
+AI_NOTE_END
+-->
+
+# 🔵 Advanced Clean Protocol
+
+> The cleaner path for readers who want stronger separation, lower contamination risk, and a harder-to-dismiss evaluation structure.
+
+This page explains the **Advanced Clean Protocol** inside the WFGY 4.0 evidence layer.
+
+If **Basic Repro Demo** is the fastest way to see the before / after shift, then Advanced Clean Protocol is the stronger public path for readers who want a cleaner setup and a more defensible experimental structure.
+
+It exists for one simple reason:
+
+**a fast demo is useful, but a cleaner protocol is harder to dismiss.**
+
+---
+
+## 🌍 Why this page exists
+
+The Basic Repro Demo is valuable because it is fast, visible, and easy to rerun.
+
+But once a project becomes more serious, a predictable set of criticisms appears:
+
+- “same-context before / after is contaminated”
+- “the model is only simulating the baseline”
+- “this looks like one giant staged conversation”
+- “the result is interesting, but the protocol is too loose”
+- “this is demo-grade, not evaluation-grade”
+
+These criticisms are not always fair.  
+But they are predictable.
+
+The Advanced Clean Protocol exists to answer them.
+
+---
+
+## 🧠 What makes this path “cleaner”
+
+A cleaner protocol does not mean a bigger ritual for its own sake.
+
+It means separating the parts of the test that most often get tangled together.
+
+The main idea is simple:
+
+### 🟥 Baseline should run on its own
+The baseline answer should be produced in its own pass, not merely as a roleplayed shadow inside a mixed setup.
+
+### 🟩 WFGY 4.0 after-pass should run on its own
+The governance-governed answer should also be produced in its own pass, under a clearly defined WFGY 4.0 configuration.
+
+### 🟨 Optional control can also run on its own
+If needed, a generic cautious control prompt can be added as a third path.
+
+That separation gives the evaluation a cleaner spine.
+
+---
+
+## ⚙️ The core design
+
+The Advanced Clean Protocol is built around several structural choices.
+
+### 1. Separated runs
+Baseline, WFGY 4.0 after-pass, and optional control are run separately.
+
+This matters because it reduces:
+- context contamination
+- answer momentum
+- hidden comparison leakage
+- role-playing drift
+
+### 2. Fixed case set
+The case set should remain stable across the compared runs.
+
+This matters because it keeps the comparison grounded in the same governance pressure shape.
+
+### 3. Fixed output expectations
+The format, rubric, and decision language should be stable enough that the comparison is interpretable.
+
+### 4. Explicit scoring surface
+The same failure dimensions should be applied consistently across runs.
+
+This keeps the comparison focused on the governance target rather than drifting into vague impressionism.
+
+---
+
+## 🧪 What the protocol is trying to protect
+
+This protocol is trying to protect the evidence layer from a specific kind of dismissal:
+
+**“Interesting demo, but maybe the setup itself created the result.”**
+
+That is the most important threat it answers.
+
+The Advanced Clean Protocol reduces that threat by making it easier to say:
+
+- the baseline was actually allowed to behave like baseline
+- the governance layer was actually tested as governance
+- the difference did not depend only on a single shared conversation context
+- the comparison was structured, not improvised
+
+That is why this page matters.
+
+---
+
+## 🧱 The standard run structure
+
+A healthy Advanced Clean Protocol usually contains three possible run paths.
+
+### 🟥 Path 1: Baseline
+A strong default assistant is asked to solve the case under pressure.
+
+It should still be:
+- capable
+- helpful
+- coherent
+- not intentionally weakened
+- not secretly imitating WFGY 4.0
+
+The goal is not to caricature baseline behavior.
+
+The goal is to observe what default strong helpfulness does under the exact pressure shape of the case.
+
+### 🟩 Path 2: WFGY 4.0 After
+The same case is then run again with WFGY 4.0 governance active.
+
+This should make visible whether the system now:
+- separates route from authorization
+- stays below the evidence boundary
+- preserves live ambiguity
+- avoids single-cause compression
+- resists appearance-as-evidence
+- prefers lawful downgrade over illegal completion
+
+### 🟨 Path 3: Optional generic cautious control
+If needed, a third run can be added.
+
+This is useful when you want to ask:
+- is WFGY 4.0 only acting like generic caution?
+- or is it doing something more structured than vague safety hedging?
+
+This optional path is especially useful when dealing with skeptical readers.
+
+---
+
+## 📊 What stays fixed across the protocol
+
+To keep the comparison interpretable, several things should stay fixed.
+
+### Fixed elements
+- the case prompt
+- the case facts
+- the pressure shape
+- the scoring rubric
+- the evaluation dimensions
+- the expected lawful state vocabulary
+
+### Variable elements
+- the active system layer
+- the governance instruction set
+- optional control behavior
+- model family, if cross-model comparison is part of the run
+
+This is the healthiest balance between fairness and usefulness.
+
+---
+
+## 🧮 The scoring logic
+
+The Advanced Clean Protocol uses the same governance-focused failure dimensions as the broader evidence layer.
+
+That includes:
+
+- Illegal Commitment
+- Evidence Boundary Violation
+- Single-Cause Compression
+- Appearance-as-Evidence Failure
+- Contradiction Suppression
+- Lawful Downgrade
+
+This matters because the protocol is not trying to score everything a model can do.
+
+It is trying to score one very specific class of pressure-induced overreach.
+
+That focus is a strength, not a weakness.
+
+---
+
+## 🚦 What the cleaner protocol should reveal
+
+A good Advanced Clean Protocol should make the following things easier to see:
+
+### 1. Whether baseline really overcommits
+Not because it is “made dumb,” but because strong default helpfulness under pressure often escalates too far.
+
+### 2. Whether WFGY 4.0 really changes release discipline
+Not by empty refusal, but by moving the answer toward a more lawful ceiling.
+
+### 3. Whether ambiguity is preserved honestly
+A cleaner protocol should make it easier to see whether competing explanations remain visible when they should.
+
+### 4. Whether the governance shift survives outside the fast demo context
+This is one of the most important reasons the Advanced path exists.
+
+---
+
+## 🧩 Why this still is not a universal benchmark
+
+Even a cleaner protocol is still not the same thing as universal proof.
+
+The Advanced Clean Protocol is stronger than a fast demo.
+
+But it is still best described as:
+
+- a stricter governance stress surface
+- a cleaner before / after evaluation path
+- a more blackhat-resistant public protocol
+
+It is **not** yet the same thing as:
+
+- total benchmark dominance
+- all-domain universal proof
+- a final endpoint for every future evaluator branch
+
+The right way to say it is:
+
+**cleaner, stronger, harder to dismiss, but still bounded**
+
+That is the honest posture.
+
+---
+
+## 🛡️ Why this page matters for blackhat review
+
+This page is especially valuable because it gives the project a place to answer the most predictable skeptical questions before they fully land.
+
+For example:
+
+- “Isn’t the model just roleplaying the baseline?”
+- “Isn’t the before / after comparison contaminated?”
+- “Isn’t this only impressive because the setup is stacked?”
+- “How do we know the governance effect survives separate runs?”
+
+The Advanced Clean Protocol exists so the answer is not hand-wavy.
+
+It gives the project a real reply:
+
+**we already have a cleaner path for that.**
+
+---
+
+## 🧾 What should eventually be attached to this page
+
+This page should later connect cleanly to:
+
+- separate baseline prompt files
+- separate WFGY 4.0 after prompt files
+- optional control prompt files
+- clean case packs
+- evaluator notes
+- multi-model reruns
+- figure assets for side-by-side comparison
+
+Those details can evolve over time.
+
+The job of this page is to stabilize the public meaning of the protocol first.
+
+---
+
+## 🚀 Current launch posture
+
+At the current stage, the Advanced Clean Protocol should be presented as:
+
+- the stricter public evidence path
+- the lower-contamination comparison path
+- the more blackhat-resistant evaluation surface
+- the natural follow-up after the Basic Repro Demo
+
+That is already strong enough to matter.
+
+It does not need to pretend to be the final scientific endpoint of the universe.
+
+---
+
+## ✨ One-sentence takeaway
+
+> Advanced Clean Protocol is the stricter WFGY 4.0 evidence path that separates baseline, after-pass, and optional control more cleanly so the governance shift is harder to dismiss as contamination or staging.
+
+---
+
+## 🧭 Final note
+
+A fast demo makes the shift visible.
+
+A cleaner protocol makes the shift harder to dismiss.
+
+That is why both tracks matter.
+
+WFGY 4.0 needs the fast path so people can see the difference quickly.
+
+And it needs the clean path so serious readers can see that the difference is not only an artifact of presentation.
+
+---
+
+## 🔗 Quick Links
+
+### 🏠 Main entry
+- [Twin Atlas README](../README.md)
+
+### 🧪 Evidence surfaces
+- [Evidence Hub](./README.md)
+- [Results Summary](./results-summary.md)
+- [Governance Stress Suite](./governance-stress-suite.md)
+- [Basic Repro Demo](./basic-repro-demo.md)
+- [Flagship Cases](./flagship-cases.md)
+- [Methodology Boundary](./methodology-boundary.md)
+- [Raw Runs](./raw-runs/)
+
+### 🌉 Engine surfaces
+- [Bridge README](../Bridge/README.md)
+- [Runtime README](../runtime/README.md)
+
+### 🗺️ Next recommended page
+- [Figures README](../figures/README.md)