Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
24 commits
Select commit Hold shift + click to select a range
6a0e3b9
reconciliation (C6 / witness collision): declarative source-bound rec…
claude Sep 29, 2026
5f11e3b
reconciliation (C6 / witness collision) v1.1: apply OpenAI 5359701805…
th3realdyll-afk Sep 29, 2026
87ba1c4
Merge Math- main 5c484bf (Math-#161 landed) into the C6 witness-colli…
claude Sep 29, 2026
3e00f21
reconciliation (C6 / witness collision) v1.2: candidate status after …
th3realdyll-afk Sep 29, 2026
da8ef02
reconciliation (C6 / witness collision) v1.3: answer Codex review 535…
th3realdyll-afk Sep 30, 2026
37470aa
Merge Math- main aa28838 (Math-#158 landed) into the C6 witness-colli…
claude Sep 30, 2026
c641df1
reconciliation (C6 / witness collision) v1.4: re-verified at main aa2…
th3realdyll-afk Sep 30, 2026
670e4ae
register alignment (D2, D3, D4, D5-annulus, D6): declarative source-b…
th3realdyll-afk Sep 30, 2026
3e17b66
Merge Math- main bb429d3 (Math-#157 and #165 landed) into the C6 witn…
claude Sep 30, 2026
5bdcfff
reconciliation (C6 / witness collision) v1.5: re-verified at main bb4…
th3realdyll-afk Sep 30, 2026
9c8a3d9
reconciliation (C6 / witness collision) v1.6: record xAI bounded disc…
th3realdyll-afk Sep 30, 2026
f2250bf
register alignment v1.1: record xAI reads 5901912476 and 5360240245; …
th3realdyll-afk Sep 30, 2026
7f51fac
Merge Math- main ec6db8c (Math-#151 D5 graph fold) into claude/regist…
claude Sep 30, 2026
55aef68
register alignment v1.2: re-verified at ec6db8c (Math-#151 merged); l…
claude Sep 30, 2026
18f7edf
Merge Math- main ec6db8c (Math-#151 D5 graph fold) into claude/c6-wit…
claude Sep 30, 2026
eba6045
reconciliation (C6 / witness collision) v1.7: re-verified at ec6db8c …
claude Sep 30, 2026
0cb128d
register alignment v1.3: evidence carried by the graph (component nod…
claude Sep 30, 2026
7ba9bbf
Refresh register alignment record onto current main
d6g8k5htny-coder Sep 30, 2026
0fc352b
Merge branch 'main' into claude/c6-witness-reconciliation-20260929
d6g8k5htny-coder Sep 30, 2026
43dd74c
Bind two-scale packet reviews to exact proof bytes
d6g8k5htny-coder Sep 30, 2026
fa1cf7b
Merge PR #167: register-alignment record
d6g8k5htny-coder Sep 30, 2026
21c7c20
Refresh two-scale review custody repair onto current main
d6g8k5htny-coder Sep 30, 2026
fe2c3ff
Merge PR #172: bind two-scale reviews to exact proof bytes
d6g8k5htny-coder Sep 30, 2026
dd4f4f8
Refresh C6 witness reconciliation record onto current main
d6g8k5htny-coder Sep 30, 2026
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
74 changes: 74 additions & 0 deletions .github/workflows/c6-witness-collision-reconciliation.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,74 @@
name: C6 witness-collision reconciliation checks

'on':
pull_request:
paths:
- 'reviews/c6_witness_collision_reconciliation_20260929/**'
- '.github/workflows/c6-witness-collision-reconciliation.yml'
- 'frontiers/c6_*/**'
- 'frontiers/d5_dimension_lift_20260929/**'
- 'frontiers/elder_dimension_lift_20260928/**'
- 'frontiers/remote_collision_20260928/**'
- 'frontiers/remote_window_20260924/**'
- 'frontiers/downstream_gate_20260925/GRAPH.json'
- 'frontiers/downstream_gate_20260925/SELECTOR_REGION.json'
- 'imports/lifetime_parent_20260925/**'
- 'reviews/d5_reconciliation_20260929/**'
- 'reviews/c6_fourier_completion_20260929/**'
- 'reviews/elder_dimension_lift_claude_20260928/**'
- 'reviews/d5_remote_collision_grok_20260928/**'
- 'reviews/d5_offpin_second_moment_20260928/**'
- 'PROOF_INDEX.md'
Comment thread
d6g8k5htny-coder marked this conversation as resolved.
Resolved
- 'reviews/candidates_pending_20260928/CANDIDATES.md'
workflow_dispatch:

permissions:
contents: read

jobs:
exact-and-probe:
runs-on: ubuntu-latest
timeout-minutes: 10
steps:
- uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262
with:
persist-credentials: false
- uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065
with:
python-version: '3.11.16'
- name: Verify sources, replay both modes, reject mutants
env:
PYTHONDONTWRITEBYTECODE: '1'
run: |
set -euo pipefail
python -B -S - <<'PY'
import hashlib, json, pathlib, subprocess, sys
root = pathlib.Path('reviews/c6_witness_collision_reconciliation_20260929').resolve()
manifest = json.loads((root / 'SOURCE_FILES.json').read_text())['files']
present = sorted(p.name for p in root.iterdir())
expected_names = sorted([e['path'] for e in manifest] + ['SOURCE_FILES.json'])
if present != expected_names:
raise ValueError('packet tree differs from manifest: ' + repr(present))
for entry in manifest:
path = root / entry['path']
if pathlib.PurePosixPath(entry['path']).name != entry['path'] or path.is_symlink():
raise ValueError('flat regular source required: ' + entry['path'])
data = path.read_bytes()
if len(data) != entry['bytes'] or hashlib.sha256(data).hexdigest() != entry['sha256']:
raise ValueError('source identity mismatch: ' + entry['path'])
script = str(root / 'reconciliation_check.py')
expected = (root / 'RESULTS.json').read_bytes()
mutants = ['allow-symlink', 'no-hash', 'drop-edge', 'stale-fingerprint', 'executed-flag', 'close-residual', 'drop-lower-bound', 'regional-strict', 'clone-live-node', 'open-cell-left']
for flags in (['-B', '-S'], ['-B', '-O', '-S']):
out = subprocess.run([sys.executable, *flags, script],
check=True, capture_output=True, timeout=300).stdout
if out != expected:
raise RuntimeError('output differs from RESULTS.json under ' + ' '.join(flags))
for m in mutants:
rc = subprocess.run([sys.executable, *flags, script, '--mutant', m],
capture_output=True, timeout=300).returncode
if rc != 1:
raise RuntimeError('mutant ' + m + ' not rejected')
print(json.dumps({'scientific_effect': 'NONE', 'mathematical_acceptance': False}))
PY
git diff --exit-code
75 changes: 75 additions & 0 deletions .github/workflows/register-alignment.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,75 @@
name: Register alignment checks

'on':
pull_request:
paths:
- 'reviews/register_alignment_20260930/**'
- '.github/workflows/register-alignment.yml'
- 'frontiers/three_fronts_20260924/LIFETIME_REMAINDER.md'
- 'frontiers/three_fronts_20260924/RN_COUNT_INTERFACE.md'
- 'coefficients/side24_v1/PROOF.md'
- 'frontiers/remote_window_20260924/PROOF.md'
- 'frontiers/rn_thin_tube_20260925/FIXED_ANNULUS_CANDIDATE.md'
- 'frontiers/full_price_20260924/PROOF.md'
- 'imports/lifetime_parent_20260925/UNIFORM_MATRIX_CAP_AND_LIFETIME.md'
- 'reviews/pr22_fixed_annulus_nonauthor_20260925/**'
- 'reviews/p15_full_price_nonauthor_20260926/**'
- 'reviews/side24_v1_coefficient_claude_20260929/**'
- 'frontiers/rn_thin_tube_20260925/TWO_SCALE_ADDENDUM.md'
- 'reviews/replacement_20260925_pr19_pr21/TWO_SCALE_REVIEW.md'
- 'frontiers/three_fronts_20260924/P15_REALIZED_COVERS.md'
- 'frontiers/downstream_gate_20260925/GRAPH.json'
- 'frontiers/downstream_gate_20260925/hard_gate.py'
- 'PROOF_INDEX.md'
workflow_dispatch:

permissions:
contents: read

jobs:
exact-and-probe:
runs-on: ubuntu-latest
timeout-minutes: 10
steps:
- uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262
with:
persist-credentials: false
- uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065
with:
python-version: '3.11.16'
- name: Verify sources, replay both modes, reject mutants
env:
PYTHONDONTWRITEBYTECODE: '1'
run: |
set -euo pipefail
python -B -S - <<'PY'
import hashlib, json, pathlib, subprocess, sys
root = pathlib.Path('reviews/register_alignment_20260930').resolve()
manifest = json.loads((root / 'SOURCE_FILES.json').read_text())['files']
present = sorted(p.name for p in root.iterdir())
expected_names = sorted([e['path'] for e in manifest] + ['SOURCE_FILES.json'])
if present != expected_names:
raise ValueError('packet tree differs from manifest: ' + repr(present))
for entry in manifest:
path = root / entry['path']
if pathlib.PurePosixPath(entry['path']).name != entry['path'] or path.is_symlink():
raise ValueError('flat regular source required: ' + entry['path'])
data = path.read_bytes()
if len(data) != entry['bytes'] or hashlib.sha256(data).hexdigest() != entry['sha256']:
raise ValueError('source identity mismatch: ' + entry['path'])
script = str(root / 'alignment_check.py')
expected = (root / 'RESULTS.json').read_bytes()
mutants = ['allow-symlink', 'no-hash', 'stale-fingerprint', 'drop-review-source', 'controlling-true', 'executed-flag', 'skip-interface-premise', 'holder-reversed', 'baseline-drift', 'duplicate-drift', 'region-unlinked', 'review-metadata-only']
for flags in (['-B', '-S'], ['-B', '-O', '-S']):
out = subprocess.run([sys.executable, *flags, script],
check=True, capture_output=True, timeout=300).stdout
if out != expected:
raise RuntimeError('output differs from RESULTS.json under ' + ' '.join(flags))
for m in mutants:
rc = subprocess.run([sys.executable, *flags, script, '--mutant', m],
capture_output=True, timeout=300).returncode
if rc != 1:
raise RuntimeError('mutant ' + m + ' not rejected')
print(json.dumps({'scientific_effect': 'NONE', 'mathematical_acceptance': False}))
PY
git diff --exit-code
5 changes: 5 additions & 0 deletions .github/workflows/two-scale-cluster-geometry.yml
Original file line number Diff line number Diff line change
@@ -1,6 +1,11 @@
name: Two-scale cluster geometry checks

'on':
push:
branches: ['main']
paths:
- 'frontiers/two_scale_cluster_geometry_20260929/**'
- '.github/workflows/two-scale-cluster-geometry.yml'
pull_request:
paths:
- 'frontiers/two_scale_cluster_geometry_20260929/**'
Expand Down
30 changes: 18 additions & 12 deletions frontiers/two_scale_cluster_geometry_20260929/MANIFEST.json
Original file line number Diff line number Diff line change
Expand Up @@ -11,9 +11,9 @@
},
{
"path": "README.md",
"bytes": 3850,
"sha256": "4e2de660b46fa3c0107a06062de42ee581bade186daac35d666bbfa4a1f95aa3",
"git_blob": "8ee02224f4a5edeec0050fb54c3cc36f85c1070b"
"bytes": 5636,
"sha256": "80e419ca23cfc4083957dcd5708fd77c82316dd65222691b54e5f30afeb11506",
"git_blob": "4036b28db7484aa85868afb9443a28f7cceedee3"
},
{
"path": "RECONNAISSANCE.md",
Expand All @@ -27,11 +27,17 @@
"sha256": "3277199a22d6f187354e6f912361648e737bc8aa65107d1d7863a099c176ca8d",
"git_blob": "99678cd670b9f8fee2f02d73fdb091705b825491"
},
{
"path": "REVIEW_RECORD.json",
"bytes": 4226,
"sha256": "958d2e1ae8ea5d034d0d87efce6cde9be57ebb9d7c0b671e69d3d6598a6675ca",
"git_blob": "a4b0fb99c24e2edf41973ea2a5e5176584602855"
},
{
"path": "REVIEW_RECORD.md",
"bytes": 5121,
"sha256": "5295c373a16b6557584e1a633e28fd5962f3e96023945bcc068819e83a0d2cad",
"git_blob": "9f7c7f6fc59ead440258887eb113135ab1dc3463"
"bytes": 6732,
"sha256": "af1540308f37220e6592b700bf42181205b1d4588a040a5c151bf0a6f861e728",
"git_blob": "31896de23fc56822c184375d0dc2b1c7e27fcae4"
},
{
"path": "SOURCES.json",
Expand Down Expand Up @@ -59,9 +65,9 @@
},
{
"path": "test_custody.py",
"bytes": 7574,
"sha256": "f46073d3ff276cd0fdc571bf76e1cda8e0e4b20d93f3f7ab52a43d936258b325",
"git_blob": "d055261a2ff56d038fcf1ce803e0167c0e04d95b"
"bytes": 13368,
"sha256": "70ff39e53dd2d763e3ddf28affddedb7468c410e95565e9d0854e393f200de31",
"git_blob": "631ebf653ea760ae6cc01afd46c23b0eb893c046"
},
{
"path": "test_geometry.py",
Expand All @@ -71,9 +77,9 @@
},
{
"path": "verify.py",
"bytes": 6002,
"sha256": "335dd00fabb70db04498a84eef926266faa9f34486458724d3b0655d10b4f4ca",
"git_blob": "e401a78a1a28200f3f7e473b9260f36919ab69f8"
"bytes": 8457,
"sha256": "152a41410fbf4774368e40900b9b988505f263d574c632fe5fcdae999a0ca6ef",
"git_blob": "6ecc60a58f8f77ee59eaa4c276e5b4ae17beedb4"
}
]
}
29 changes: 29 additions & 0 deletions frontiers/two_scale_cluster_geometry_20260929/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -65,3 +65,32 @@ source identity to be a commit object, not merely a forty-hex tree-ish name. Fou
real-Git red/green regressions cover forged/legitimate replacement bindings and
tree/tag substitution. Both mathematical manuscripts and all mathematical inputs
remain byte-identical to the initial reviewed head.

## Current proof-to-review custody replay

The earlier 52-test count describes the preceding custody repair. The current
replay runs 61 unit tests in each Python mode, including nine additional test
methods for proof and review bindings. The verifier requires both manuscripts,
the source/result declarations, both review records, the executable checker and
verifier, and both test modules. Regenerating MANIFEST.json cannot silently omit
these files or move an old review to different proof bytes.

REVIEW_RECORD.json records the actual A/B/C native IDs, COMMENTED states, raw
review-body byte counts and SHA256 digests, exact reviewed commit, proof mapping,
conditional scopes and exposure. The verifier anchors the entire frozen record
and the accepted manuscript/source identities outside the regenerable inventory.
Intentional future mathematical changes require reviewed updates to those
anchors and the review record. Both current manuscripts, SOURCES.json,
RESULTS.json and geometry.py remain unchanged.

This is an offline consistency check. Its successful summary explicitly reports
remote_review_authenticity_checked=false and mathematical_acceptance=false.
The native bodies were actually fetched during integration preflight; the local
CLI does not fetch them or authenticate a remote reviewer. In --local-only mode,
the source declaration's frozen bytes are still checked, while the historical
Git objects remain excluded and source_pins_checked=false.

The same workflow now also runs for main pushes affecting this packet or its own
workflow, providing a scoped post-merge replay route. Prior PR/manual coverage
was not itself a verifier defect. Hosted success must bind the actual candidate
or merge commit; this local note supplies no CI result or mathematical vote.
Original file line number Diff line number Diff line change
@@ -0,0 +1,77 @@
{
"schema": "math166-scoped-review-binding-v1",
"scientific_effect": "NONE",
"mathematical_acceptance": false,
"remote_review_authenticity_checked": false,
"reviewed_commit": "7c82252533c3fe14ee262f7f87c0549c10968592",
"proofs": [
{
"path": "TWO_SCALE_LAW.md",
"bytes": 17139,
"sha256": "e81d7fe09d25c3266eb8e62922756f54761d7f74d692fb35d9a8071fffd73769",
"git_blob": "a32fd5f7d941bbe1fe943df045b1e0fbec8d691c"
},
{
"path": "RADIAL_TAIL.md",
"bytes": 12533,
"sha256": "250897858c8b314c0ff85ec1860a725efacfdae0ff7409fced4cadf7bfb85973",
"git_blob": "0f14417ef7038b9c6f50e4b01e393a58b7b5e3a4"
}
],
"source_manifest": {
"path": "SOURCES.json",
"bytes": 2923,
"sha256": "de09c2cf3c9082a09d168f451f78a1a60b150f9f55c0c6b9bfe94ac26652dfe0",
"git_blob": "59d11113e208965fefeaadd1bd5866585e77dc82"
},
"reviews": [
{
"slice": "A",
"native_review_id": 5360227991,
"native_state": "COMMENTED",
"url": "https://github.com/d6g8k5htny-coder/Math-/pull/166#pullrequestreview-5360227991",
"submitted_at": "2026-09-30T00:52:00Z",
"reviewed_commit": "7c82252533c3fe14ee262f7f87c0549c10968592",
"proof_path": "TWO_SCALE_LAW.md",
"body_bytes": 7840,
"body_sha256": "c63efc292d7cf209e2770419cc159276897d0ba955c32ee1d2af88c8aa7f0ed8",
"reported_disposition": "ACCEPT at the stated conditional scope",
"scope": "Theorem T, Theorem L, ordered-pair law L3, sampling/order of limits and spatial-TV obstruction V1; conditional on exact imported source interfaces, without a rate or parent revalidation."
},
{
"slice": "B",
"native_review_id": 5360178611,
"native_state": "COMMENTED",
"url": "https://github.com/d6g8k5htny-coder/Math-/pull/166#pullrequestreview-5360178611",
"submitted_at": "2026-09-30T00:42:15Z",
"reviewed_commit": "7c82252533c3fe14ee262f7f87c0549c10968592",
"proof_path": "RADIAL_TAIL.md",
"body_bytes": 7631,
"body_sha256": "3042d958b769b0a9ada1359b6f8d9461c090060fe5057cff795c99b6c6c2894d",
"reported_disposition": "ACCEPT at the stated conditional scope",
"scope": "RADIAL_TAIL R4-R13: the already-formed limiting microscopic point-intensity tail, full coefficient and moment boundary; conditional on SC/CUB, with r tending to zero before radius tends to infinity."
},
{
"slice": "C",
"native_review_id": 5360216551,
"native_state": "COMMENTED",
"url": "https://github.com/d6g8k5htny-coder/Math-/pull/166#pullrequestreview-5360216551",
"submitted_at": "2026-09-30T00:49:42Z",
"reviewed_commit": "7c82252533c3fe14ee262f7f87c0549c10968592",
"proof_path": "RADIAL_TAIL.md",
"body_bytes": 4277,
"body_sha256": "9a2c89c6e0c122fc75cc96a2b6392b47ab33bc47feb072b9f2a56315c527df84",
"reported_disposition": "ACCEPT at the stated conditional scope",
"scope": "RADIAL_TAIL H1-H4: extreme marks under point-intensity sampling of the already-formed microscopic measure, not a uniformly sampled configuration or finite-r global law.",
"proof_binding_note": "The native C review refers to B's same full Git blob and an abbreviated SHA256. The integration preflight matched the complete proof bytes at the native reviewed commit; this record does not claim C printed a fresh full digest."
}
],
"exposure": {
"reviewer_provider": "Anthropic",
"reviewer_model_attribution": "Claude",
"same_session_for_all_three": true,
"organizational_independence_credit": 0,
"notes": "Actual native review bodies were read back by the integration preflight. These are three disjoint scopes from one provider-distinct session, not three independent votes. Parent proofs remain imported; no new analytic review is supplied by this custody repair."
},
"binding_policy": "The verifier pins this complete record and the accepted proof/source byte identities. Regenerating MANIFEST.json cannot transfer these old reviews to changed mathematics. An intentional future update requires reviewed changes to the bindings and verifier anchors. Offline replay does not contact GitHub or authenticate a remote reviewer."
}
25 changes: 25 additions & 0 deletions frontiers/two_scale_cluster_geometry_20260929/REVIEW_RECORD.md
Original file line number Diff line number Diff line change
Expand Up @@ -90,3 +90,28 @@ source run36652196578 and full downstream/formal run36652196938 both succeeded
at the repair commit. A later metadata successor still needs fresh current-head
checks; no old-head run is automatically rebound. The mathematical files,
geometry.py, source pins and RESULTS.json have not changed.

## Current integration custody repair

At b8b028ea0a3f45aa8d2640b339977a04154909a4, independent full-CLI fixtures showed
that changing either manuscript, or removing either manuscript, still passed
after honestly regenerating MANIFEST.json. The ordinary unresigned mutation
correctly failed. These were inventory/review-binding defects, not mathematical
counterexamples. The original reviewed manuscripts remain unchanged.

The current repair requires the proof/review declarations and every executable
or test module used by the advertised replay. It binds both current manuscripts
and SOURCES.json to fixed accepted identities and pins the complete additive
REVIEW_RECORD.json, including actual native review body identities. A co-edit of
a manuscript, review metadata and manifest cannot rebind an old review merely by
re-signing the mutable inventory. Nine new regression methods cover these cases;
the current suite contains 61 tests per Python mode. Existing exact geometry,
eight semantic mutants and Git replacement/commit-type controls remain intact.

The A/B/C records above retain their original conditional scope and exposure.
C's native proof identity cross-references B's full blob and gives an abbreviated
digest; actual bytes at the native reviewed commit were independently matched
in preflight. The machine-readable record makes that distinction explicit.
Offline validation authenticates local byte relationships only: it grants no
remote-review authentication, mathematical acceptance or independence credit.
Fresh engineering review, current-head CI and integration remain separate gates.
Loading
Loading