peer-review: align validation language, add Python 3.11/3.12 CI matrix

- VALIDATION.md: replace 'the physics is correct' with scoped statement
  referencing LIMITATIONS.md for evidence boundary
- validate_all.py: replace 'trust these results' with 'internally consistent
  and agrees with published reference data'
- ci.yml: expand test matrix to Python 3.10, 3.11, 3.12 (matches
  pyproject.toml classifiers)
- GitHub Release: deleted orphaned draft release (sidebar now shows v0.1.0)

Addresses reviewer feedback items 2, 3, 4 from latest audit.

Author: Aethermor

.github/workflows/ci.yml

@@ -9,14 +9,17 @@ jobs:
   test:
     runs-on: ubuntu-latest
     timeout-minutes: 45
+    strategy:
+      matrix:
+        python-version: ["3.10", "3.11", "3.12"]
     steps:
       - name: Checkout
         uses: actions/checkout@v4
 
-      - name: Set up Python
+      - name: Set up Python ${{ matrix.python-version }}
         uses: actions/setup-python@v5
         with:
-          python-version: "3.10"
+          python-version: ${{ matrix.python-version }}
 
       - name: Install dependencies
         run: |

VALIDATION.md

@@ -342,13 +342,17 @@ require proprietary floorplan data and direct thermal measurement).
 ## Interpretation
 
 - **133/133 PASS** (physics) + **20/20 PASS** (literature) + **33/33 PASS**
-  (real-world chips) means every model agrees with published data, analytical
-  solutions, conservation laws, and real chip specifications.
+  (real-world chips) means every model agrees with published reference data,
+  analytical solutions, conservation laws, and real chip specifications
+  within the scope of the tests listed above.
 
 - **Any FAIL** means something is wrong.  The failure message tells you
   exactly which model, what was expected, and what was observed.  File an
   issue or investigate before using results from that model.
 
-- **The validation suite is the contract.**  If it passes on your machine,
-  the physics is correct.  If it fails, something in the environment or
-  code has changed and needs investigation.
+- **The validation suite is a necessary but not sufficient check.**  If it
+  passes on your machine, the models are internally consistent and agree
+  with the published references listed above.  This does not substitute
+  for validation against direct hardware measurement or proprietary
+  simulation tools, which remains outside the current scope
+  (see [LIMITATIONS.md](LIMITATIONS.md)).

validation/validate_all.py

@@ -18,8 +18,8 @@
   5. Constraint satisfaction     (optimizer respects its own contracts)
   6. Monotonicity & limit checks (correct qualitative behaviour)
 
-If every check reads PASS, a researcher can trust that Aethermor's
-predictions are grounded in validated physics — not just code that runs.
+If every check reads PASS, the models are internally consistent and
+agree with the published references listed in each section.
 
 Exit code: 0 if all checks pass, 1 if any fail.
 """
@@ -884,9 +884,9 @@ def main():
         print()
         print("  ✓ ALL CHECKS PASSED")
         print()
-        print("  Every model has been validated against published reference data,")
-        print("  analytical solutions, conservation laws, and internal consistency.")
-        print("  Researchers can trust these results.")
+        print("  Every model is internally consistent and agrees with published")
+        print("  reference data, analytical solutions, and conservation laws.")
+        print("  See LIMITATIONS.md for scope of validation.")
         print()
     else:
         print()