[K3] Close c112 source-line and raw 433-1b/O0a workboards

[AllenGrahamHart]

Summary

Living progress PR for upstream workboard item K3: KoalaBear MCA at 2^-128, order-two m=2,r=4 class.

Nine certificate packets now live under experimental/data/certificates/:

• c112 source-line coverage COMPLETE: all aligned-positive, near-positive, aligned-negative, near-negative, projective-boundary, and M01-R11/M02-R11 source-line cells.
• 433-1b role-cell closures: cells 3, 4, 9, 11, and 12-13.
• 433-1b role orbit [5,8] CLOSED by exact B/C transport.
• 433-1b universal structure: the 105-label quotient, common compilers, and role transports.
• 433-1b -> O0a raw workboard COMPLETE: the exact owner partition
  [0]|[1,2]|[3,6]|[4,7]|[5,8]|[9,10]|[11]|[12,13]|[14]
  covers all 15 role cells, 1,575 labels, and 25,200 signed principal systems; the global rank-drop theorem excludes the complement.

The latest packet repairs an implicit step in the earlier export: it proves an exact B/C duplicate-role bijection from cell 9 to cell 10 over all 105 labels and 1,680 signed principal systems. Its self-verifier checks ten blobs at the pinned public campaign commit.

Exact provenance and replay

The aggregate packet pins AllenGrahamHart/rs-mca-prize-dag commit
8df0903391a228eed6e24398fca9d40d72d546cf.

python3 experimental/data/certificates/kb-mca-v4-433-1b-raw-workboard-close-v1/verify.py \
  --source-root /path/to/rs-mca-prize-dag

Expected terminal:

RAW_WORKBOARD_CERT_PASS nodes=3 hashes=10 cells=15 labels=1575 systems=25200

Scope and remaining K3 frontier

This PR does not close K3 or the KoalaBear row, and moves no ledger value.

The closed object is the raw 433-1b -> O0a route. The exact positive-coordinate workboard has thirteen routes; only 433-1a -> O0b and 433-1b -> O0a are closed, leaving eleven route payments. The broader K3 proof still needs:

• exhaustive same-owner orientation/source-cover assembly;
• literal distinct-affine-slope payments for every survivor;
• genuine-pencil hypotheses and exact relevant-line counts for nonzero pencil payments;
• exact ray compilers for higher-dimensional survivors;
• the active K3 allocation inequality;
• fresh independent mathematical and custody review.

Raw labels and signed principal systems are not being identified with distinct bad slopes.

Credit

Continues the #1122-#1155 K3 series. Scott Hughes PRs #1153-#1155 supply useful independent replay and the cell-11 guard-transplant route cut; those packets do not conflict with this stronger raw-cell closure.

[scottdhughes]

Independent replay blocker for the new [5,8] unit: the certificate pins public-DAG commit 3fa2987430242cb631ab76be4ebbee549ce95fb8, but as of 2026-08-09T19:30:12Z direct fetch returned upload-pack: not our ref, the GitHub commit API returned HTTP 422 / No commit found, and the named branch codex/full-prize-resolution-v12-20260807 was absent from git ls-remote. The load-bearing pairing-7/8/11, cell-5 aggregate, [5,8] aggregate, and duplicate-role transport nodes were also absent from current public master fc7133243.

This is not a mathematical counterexample, but it blocks the advertised 23-primary + 23-audit independent replay and therefore blocks a GREEN provenance audit. Please push the exact commit or re-pin the certificate to a durable public commit. Once available, I will replay all 46 scripts and audit endpoint rootlessness plus duplicate-role transport. PR #1153 now provides an independent public-source replay of the six residual xi3 cell-5 pairing representatives, but explicitly does not substitute for those endpoint/transport nodes.

[AllenGrahamHart]

Thanks — you checked during the known window between the worker's certificate push and the coordinator's audited merge. As of public master 711fcb9775ef6561fde1f6e7eb11d28e0a04d205 (2026-08-10), the pinned commit 3fa2987430242cb631ab76be4ebbee549ce95fb8 is reachable from master, and all the nodes you listed (the pairing-7/8/11 packets, the cell-5 aggregate, the [5,8] aggregate, and the duplicate-role transport) are on master with their verifiers.

The branch named in the provenance block is worker-local; the commit hash is the durable pin — we are annotating the certificates accordingly in a follow-up commit on this PR so the next replay doesn't trip on it.

Coordinator-side replay status: 141/141 verifier runs green at the audited pin, including every node above. Your 46-script replay would be very welcome — and #1153 is noted with thanks; we'll review it for import.