d6g8k5htny-coder · Oct 1, 2026 · Sep 30, 2026 · Oct 1, 2026 · Oct 1, 2026 · Oct 1, 2026
@@ -0,0 +1,45 @@
+name: D1 external cap audit controls
+
+'on':
+  pull_request:
+    paths:
+      - 'reviews/d1_external_audit_20260930/**'
+      - '.github/workflows/d1-external-cap-audit.yml'
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+jobs:
+  exact-and-source:
+    runs-on: ubuntu-latest
+    timeout-minutes: 10
+    steps:
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262
+        with:
+          persist-credentials: false
+          fetch-depth: 0
+      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065
+        with:
+          python-version: '3.11.16'
+      - name: Authenticate the frozen source set
+        env:
+          GIT_NO_REPLACE_OBJECTS: '1'
+        run: |
+          set -euo pipefail
+          git fetch --no-tags origin 3e0a91b69892b9717564c592e7a966492a81566a
+          python -B -S reviews/d1_external_audit_20260930/verify_sources.py
+      - name: Exact finite controls in both interpreter modes
+        env:
+          PYTHONDONTWRITEBYTECODE: '1'
+          GIT_NO_REPLACE_OBJECTS: '1'
+        run: |
+          set -euo pipefail
+          cd reviews/d1_external_audit_20260930
+          python -B -S -m unittest -v
+          python -B -O -S -m unittest -v
+          python -B -S cap_audit.py > "$RUNNER_TEMP/cap-normal.json"
+          python -B -O -S cap_audit.py > "$RUNNER_TEMP/cap-optimized.json"
+          cmp "$RUNNER_TEMP/cap-normal.json" "$RUNNER_TEMP/cap-optimized.json"
+          python -B -S verify_sources.py
+          git diff --exit-code
@@ -0,0 +1,54 @@
+{
+  "scientific_effect": "NONE",
+  "new_analytic_review": "REQUESTED_NOT_DELIVERED",
+  "files": [
+    {
+      "path": "AUDIT.md",
+      "bytes": 19468,
+      "sha256": "98356c5625aa07bcf62b9e6f1a3869ec1fea7dd789e8ea495a97b5c7abb473cf",
+      "git_blob": "fdfccaf5356f4cc0cf8764bf323f94b6ff071768"
+    },
+    {
+      "path": "README.md",
+      "bytes": 3014,
+      "sha256": "f7d91014a4c0e5e7b54f27a845c67f9dec1e6dca89609fb4ce6413dbf90fe117",
+      "git_blob": "efbfabca522ec44fafe5d0092a56ad653a429ed1"
+    },
+    {
+      "path": "SOURCES.json",
+      "bytes": 3513,
+      "sha256": "46f085619ff7c055bbd48c0381445ee3d0e01d2467c3d2586b7ac0e51262df8e",
+      "git_blob": "1f6296b20dc4d53d81be1ad54e47617f8f49013c"
+    },
+    {
+      "path": "VALIDATION.md",
+      "bytes": 2555,
+      "sha256": "7010fa762f5be2989cac7e5297e560dcc791aeb20611018cfc30bbddae858cbc",
+      "git_blob": "d83b1261201f0adceb80abd154d4f5b7ef8778e7"
+    },
+    {
+      "path": "cap_audit.py",
+      "bytes": 6143,
+      "sha256": "0a667719fef1e19ea45e3a93e493d56181820e111156d24b138fb60ca5cc6638",
+      "git_blob": "79bc4ba0c94fe5eebc6f012fff15c8107d1bc867"
+    },
+    {
+      "path": "test_cap_audit.py",
+      "bytes": 6745,
+      "sha256": "9a0893d12590e306efa4271c31e66ead01274a9f5ec37c7d90eaa65edc2a5c4d",
+      "git_blob": "73110a5f1b7e78b94a3f403e62e6a7d99d6d4a55"
+    },
+    {
+      "path": "test_sources.py",
+      "bytes": 3441,
+      "sha256": "a265ddd9c13fb9ab7d1dcaaf60de8eefbfe834b80f7f99eacc257b6b79a3aec9",
+      "git_blob": "76850b64223186a9e4894f7c33d541e8a7f0b7d6"
+    },
+    {
+      "path": "verify_sources.py",
+      "bytes": 4441,
+      "sha256": "6acf67266b0cb4328364658b0c67ea489cc2fb046c2fdd7f0668c9bad439aba5",
+      "git_blob": "2a8606478bbb53f1d18728cecfca851da25ff15f"
+    }
+  ]
+}
@@ -0,0 +1,57 @@
+# D1 cap audit and response to the external critique
+
+**Read `AUDIT.md`.** It reconstructs the deterministic cap-to-global-elder
+implication with every boundary face and the maximin argument, derives invariance
+under arbitrary exterior modifications, and gives an exact polynomial counterexample
+to replacing the whole-cylinder assumptions with endpoint jets.
+
+Status: source/author-exposed OpenAI / GPT-6 Astra Pro reconstruction, with new
+nonauthor review requested. Not an independent human/institutional assessment.
+Scientific effect NONE. No parent proof, erratum, scientific-status register,
+formal scope or SIDE24 coefficient was changed. The current user instruction
+specifically authorizes this foreground audit; it restarts no background process.
+Coordination and unresolved validation requests: Math- issue #193.
+
+The response separates:
+
+- a complete deterministic proof of the cap implication from the Gaussian/Kac-Rice
+  interfaces it does not independently re-prove;
+- D1's leading theorem from the separate reviewed D2 O(1) unrestricted remainder;
+- a compact O(ell^(2/3)) difference from an unrestricted difference;
+- an arithmetic coefficient enclosure from finite-lifetime approximation error;
+- recorded AI reviews from external-human review and kernel formalization;
+- the merged, declarative C6 residual result from a scientific register transition.
+
+## Reproduce the finite controls
+
+From the packet directory, using standard-library Python:
+
+    python -B -S -m unittest -v
+    python -B -O -S -m unittest -v
+    python -B -S verify_sources.py --local-only
+
+The last command verifies this packet's bytes and explicitly does NOT authenticate
+upstream historical Git paths. In a full project checkout use instead:
+
+    python -B -S reviews/d1_external_audit_20260930/verify_sources.py
+
+The target workflow performs that actual ten-source check. Six entries also have
+supplied byte-size and SHA256 identities; four contextual records are bound by
+historical path and Git blob only. Those differences are explicit in SOURCES.json.
+The historical reads disable Git replacements and reject noncommit/symlink inputs.
+The unit fixtures verify that behavior on temporary repositories, not on fake
+copies of project sources.
+
+The 21 tests include exact rational margins, vector derivative identities, the
+congruence erratum, the full polynomial shortcut inequality and two independent
+finite-graph algorithms over 1,728 exterior variants. They are NOT an analytic
+proof of the Gaussian statements, a simulation of persistence bars or a Lean
+formalization. The written deterministic argument, not the tests, resolves the
+logical cap question.
+
+## External-review boundary
+
+AUDIT §9 gives a focused referee checklist. Human peer review and full D1 kernel
+verification remain unsupplied. Neither an issue closure, AI label, hash nor green
+workflow is used to manufacture them. Actual hosted results and any new review
+verdict belong in the PR discussion after readback; no success is anticipated here.
@@ -0,0 +1,79 @@
+{
+  "repository": "d6g8k5htny-coder/Math-",
+  "commit": "3e0a91b69892b9717564c592e7a966492a81566a",
+  "sources": [
+    {
+      "id": "P",
+      "path": "imports/lifetime_parent_20260925/UNIFORM_MATRIX_CAP_AND_LIFETIME.md",
+      "git_blob": "dfed3b8d318a3ab1950957f393307733a4bef3f2",
+      "use": "Definitions and G_r; Gaussian/lifetime claims are not independently re-proved here.",
+      "bytes": 40261,
+      "sha256": "9350ad6eaba6626b93c3dedeef9e2ff816e5cdf1c8318e85fb27499141c84bc7"
+    },
+    {
+      "id": "CAP",
+      "path": "imports/lifetime_parent_20260925/MARKED_CYLINDER_CAP_PROOF.md",
+      "git_blob": "0633aca3c2a2882b0de4399da0a75d64c2e6b2e1",
+      "use": "Sections1-5 deterministic theorem reconstructed in full; numerical probability section not consumed.",
+      "bytes": 15160,
+      "sha256": "0bf922b9203c29088b12388807aa0e2ecd020485eb0f6e919679841b5b2636fc"
+    },
+    {
+      "id": "E1",
+      "path": "imports/lifetime_parent_20260925/ERRATUM_CONGRUENCE.md",
+      "git_blob": "213594d6ca6a86fb938110f4d166d9ce275a02d0",
+      "use": "Mandatory corrected congruence.",
+      "bytes": 1782,
+      "sha256": "bad7ef609c4ad8c41ad6af562c1b6807921e19a9d556ed793ad1a0db6e202028"
+    },
+    {
+      "id": "E2",
+      "path": "reviews/d1_section9_borel_repair_20260925/REPAIR.md",
+      "git_blob": "fe9b9ce4999908bb3814b500ee2d0ceb0c6f704a",
+      "use": "Mandatory parent Kac-Rice reading rule; not independently audited in this cap-only pass.",
+      "bytes": 9062,
+      "sha256": "845abf9f9c99d672c2a10a887b5a2e7206a3d2de3d876f35f75ff6e2dc13e62f"
+    },
+    {
+      "id": "REC",
+      "path": "reviews/d1_chain_reconciliation_20260928/RECONCILIATION.md",
+      "git_blob": "75da2597971510f843f8d90c743950cb8c177342",
+      "use": "W1, embedding radius, existing recorded review scopes; not human-review evidence.",
+      "bytes": 23312,
+      "sha256": "451b9d7ffee072a73fc904cab891b6693e3f233808b89df10cce1b57036b65da"
+    },
+    {
+      "id": "R",
+      "path": "frontiers/three_fronts_20260924/LIFETIME_REMAINDER.md",
+      "git_blob": "247b3ecf80bfbe896948d5d489b2d5842a81c481",
+      "use": "Context: TheoremR O(1) remainder; not re-proved here.",
+      "bytes": 17734,
+      "sha256": "380b7d0abdb0fe2de5a6564565af9560d3f1b1ce22fd5538927db0c52f4f3a4a"
+    },
+    {
+      "id": "R_REVIEW",
+      "path": "reviews/d2_remainder_full_depth_claude_20260930/REVIEW.md",
+      "git_blob": "529c5264ed790ab1df36c14f567f460e2a5ed974",
+      "use": "Context: recorded R1-R6 review and its exposure, not a fresh verdict by this executor."
+    },
+    {
+      "id": "SIDE24",
+      "path": "coefficients/side24_v1/PROOF.md",
+      "git_blob": "44b66f04f89fcd87383b3603fa69f1feb64cdddd",
+      "use": "Context: expression-only enclosure, cone calculation and exact covariance scope. Full interval program not rerun."
+    },
+    {
+      "id": "REGION",
+      "path": "reviews/c6_residual_closure_20260930/RECONCILIATION.md",
+      "git_blob": "3cfd17c5ef0356dbd3e8f98a2d08cb404ddd982c",
+      "use": "Context: leading-mass-localization residual, declarative only, not a D1 premise."
+    },
+    {
+      "id": "FORMAL",
+      "path": "formal/SCOPE.md",
+      "git_blob": "e5094ac386b72668b7dc42805bf8e6e64a731fac",
+      "use": "Context: exact scalar pilot scope, not full D1 formalization."
+    }
+  ],
+  "binding_note": "All ten historical Git paths/blobs checked by the hosted verifier. Six have an additional supplied size/SHA256 binding; none is invented for the other four context records."
+}
@@ -0,0 +1,43 @@
+# Actual execution record
+
+Source-exposed author-side execution only; not a new nonauthor mathematical verdict.
+
+The 15-test core was written before implementation. The deliberately invalid
+placeholder produced 13 failed assertions and one error; one test happened to
+pass on the placeholder. After implementation all 15 passed. The six-method
+real-Git source-verifier placeholder produced 13 failed assertions. After
+implementation all six passed, for 21 tests total.
+
+The final applicable run executes the complete 21-test suite in normal Python
+and with -O, compares the rational constants' serialized output in both modes,
+and checks the packet manifest. Direct upstream network retrieval failed DNS;
+local source-object authentication is therefore NOT claimed. The hosted workflow
+must authenticate the ten historical commit/path/blob bindings, plus the six
+additional declared SHA256/size bindings, before that result may be reported.
+
+Core test coverage:
+- exact positive Hermite kernel and mark rescaling;
+- strict depth threshold and fourth-derivative cutoff;
+- vector-average example with no simultaneous Rolle zero;
+- exact full vector ridge third derivative on nine polynomial ridge configurations;
+- corrected versus wrong endpoint Hessian congruence;
+- all endpoint partial derivatives through degree3 in the finite-jet falsifier;
+- polynomial-square identity bounding the entire older-reaching shortcut;
+- failure of the fourth-derivative good event for that falsifier;
+- path-only, separator-only and equal-birth finite-graph negative examples;
+- 1,728 exterior graph modifications, checked by independent widest-path and
+  threshold-connectivity algorithms.
+
+Custody fixtures include nested working directories, wrong identities, paths and
+inventory membership, noncommit objects, symlinks and real Git replacement refs.
+No actual source tampering was observed or asserted. These are temporary fixtures.
+
+The SIDE24 interval program and the complete D1 Gaussian argument were not rerun
+or formally verified. Their stated results are source comparisons in AUDIT §8.
+No local Lean/lake executable was available. Existing formal pilot execution, if
+green, retains its narrower scalar scope and cannot establish D1.
+
+One narrow external search in this pass located the primary Armentano–Azais–Leon
+arXiv record2304.07424 (the Kac-Rice source named in P). Access was at search/
+abstract depth only; no fresh theorem-level literature review is claimed. The
+cap argument is reconstructed directly and does not use that search as a premise.