From b20796620022867d06f49e048bf27b60981ac70d Mon Sep 17 00:00:00 2001
From: Mayckon Giovani <doomhammerhell@gmail.com>
Date: Wed, 23 Sep 2026 16:02:06 -0300
Subject: [PATCH] docs: prepare v0.2.0 release notes

---
 CHANGELOG.md            | 10 ++++++++++
 RELEASE_NOTES_v0.2.0.md | 33 +++++++++++++++++++++++++++++++++
 scripts/run_audit.sh    |  2 +-
 3 files changed, 44 insertions(+), 1 deletion(-)
 create mode 100644 RELEASE_NOTES_v0.2.0.md

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 64b4bc8..2775ec6 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,5 +1,15 @@
 # Changelog
 
+## 0.2.0 — External Solver Boundary
+
+- Added canonical QUBO exchange parity checks across internal models, canonicalized models, structured exchange payloads, and optional local Ocean BQM energy reports.
+- Added no-credential Ocean and Qiskit examples that enrich local reports when optional packages are installed while preserving deterministic local authority.
+- Added external solver replay artifacts with schema `noetheris.external_solver_replay.v1`, structured rejection reasons, artifact hashes, assignment-domain checks, and solver/embedding metadata boundaries.
+- Added certificate fail-closed validation for inconsistent embedded external replay evidence.
+- Added deterministic solver-boundary release evidence under `docs/results/solver_boundary_evidence.json`.
+- Added Lean and TLA+ replay-boundary formal artifacts scoped to simplified kernels.
+- Hardened Python runner selection for macOS universal Python environments.
+
 ## 0.1.0 — Structural Quantum Security Kernel
 
 - Added Structural IR in Rust and Python with canonical hashing and validation.
diff --git a/RELEASE_NOTES_v0.2.0.md b/RELEASE_NOTES_v0.2.0.md
new file mode 100644
index 0000000..3fba7b8
--- /dev/null
+++ b/RELEASE_NOTES_v0.2.0.md
@@ -0,0 +1,33 @@
+## Noetheris v0.2.0 — External Solver Boundary
+
+Noetheris v0.2.0 is a narrow research-kernel release focused on the boundary between deterministic local models and optional external solver ecosystems.
+
+The release strengthens what Noetheris exports, what an external tool may return, and what Noetheris verifies locally before treating a solver candidate as evidence.
+
+### Included
+
+- Canonical QUBO exchange parity across internal QUBO evaluation, canonicalized models, structured exchange payloads, and optional local Ocean `dimod.BinaryQuadraticModel` energy checks.
+- Local Ocean BQM reporting that runs without D-Wave credentials, sampler calls, cloud tokens, or hardware submission.
+- Qiskit-facing oracle semantic reports for small predicates, with exact local truth-table authority and optional local circuit summaries when Qiskit is installed.
+- External solver replay artifacts using schema `noetheris.external_solver_replay.v1`.
+- Explicit replay rejection reasons for problem-hash mismatch, compiled-model-hash mismatch, missing variables, unknown variables, reported-energy mismatch, malformed solver metadata, and malformed embedding metadata.
+- Certificate validation that fails closed when embedded external replay evidence is rejected, inconsistent, or artifact-hash invalid.
+- Deterministic solver-boundary evidence under `docs/results/solver_boundary_evidence.json`.
+- Lean replay-boundary kernel and TLA+ single-candidate replay state sketch, scoped to simplified formal predicates.
+- macOS Python runner selection hardened for universal Python environments where shell architecture and installed NumPy architecture can differ.
+
+### Positioning
+
+Noetheris v0.2.0 does not execute IBM Quantum or D-Wave hardware, does not benchmark cloud solvers, does not infer hardware embedding quality, and does not claim quantum advantage.
+
+External solver metadata and embedding metadata are recorded as supplied evidence. Local replay remains the verification authority: hashes, assignment domains, and energies must verify locally before a candidate can be accepted.
+
+### Validation
+
+The release gate is:
+
+```bash
+bash scripts/run_audit.sh
+```
+
+The audit covers Rust check/test/fmt/clippy, Python compilation and pytest, examples, CLI checks, deterministic release evidence generation, benchmarks, optional-backend boundary checks, Lean build when available, and release-residue source scans.
diff --git a/scripts/run_audit.sh b/scripts/run_audit.sh
index 9c55476..ef66b85 100755
--- a/scripts/run_audit.sh
+++ b/scripts/run_audit.sh
@@ -52,7 +52,7 @@ hex_terms=(
 )
 pattern="$(for encoded in "${hex_terms[@]}"; do printf '%s\n' "${encoded}" | xxd -r -p; done | paste -sd '|' -)"
 if rg -n -i "${pattern}" \
-  README.md RELEASE_NOTES_v0.1.0.md SECURITY.md CONTRIBUTING.md CODE_OF_CONDUCT.md ROADMAP.md CHANGELOG.md \
+  README.md RELEASE_NOTES_v0.1.0.md RELEASE_NOTES_v0.2.0.md SECURITY.md CONTRIBUTING.md CODE_OF_CONDUCT.md ROADMAP.md CHANGELOG.md \
   docs python crates examples tests formal scripts benchmarks .github; then
   echo "audit phrase scan failed" >&2
   exit 1