[thakrarsagar]

This finite bit-circuit construction saves 20 scalar XORs in the merged PR589 punctured R439 word. Two phase-A mini-nodes go from 20 to 14 and 13 to 9 XORs; their complete literal reverse cleanup saves ten more. Including scatter, the native word goes from 24,305 to 24,285 XORs.

The contribution exhausts all 186 destination-basis factorizations for the four two- or three-destination mini-nodes. The minima are restricted to preparation, read-only external injection and recombination, with no new wires or external mutations. All 661 dirty-input columns, every register's frame path, 439 helpers, nine retained copy ranks of seven and the paid profile are preserved. Frames 507 and 508 have no improvement in this class.

No new bit exponent or kappa is claimed. Cancellation-free row norms increase on some destinations; hosted recursive coefficient bills have not been recertified. This limitation and its exact witnesses are recorded alongside the reduction.

Base: 66849e1475c930fd7732b1a0067ffee7e1e95f60. Assignment seed: 8, OS randomness, bit-supplier track. Sources are pinned in research/p111-mini-node-basis/certificate.json. No selected construction, upstream source or existing generated artifact changes.

Reproduce from the repository root:

python3 -B research/p111-mini-node-basis/verify.py

Expected: PASS_P111_MINI_NODE_BASIS, 186 candidates, 20 XORs saved, 661 columns, unchanged frame paths/profile, independent interface acceptance and four negative controls.

Validation completed:

• Local exact synthesis and dirty-column replay, also under Python isolated mode.
• Unchanged PR555/PR589 independent E8 checker: 636 closed frames, all H1-H5 hypotheses and all U/J/K/C conclusions pass.
• Controls reject omitted preparation, omitted cleanup and illegal frame descent. A stale-cleanup control passes the endpoint test but fails the literal-reversal requirement.
• Separate agent review independently enumerated the minima and checked all 2,048 and 512 local Boolean inputs of the changed nodes.
• make selected-record-check and git diff --cached --check pass.

The repository-wide make verify, expanded hosting replays and recursive coefficient bills were not run. This isolated local contribution does not modify the selected supplier or its certificate.

Credit: eumemic's PR543/PR555; DreamingOfClouds' PR545/PR510/PR589, with the retained upstream notices. Concurrent PR670 and PR713 are credited as related finite-circuit work. This result adds different, punctured-word mini-nodes and an exhaustive small-group comparison. The final overlap check also inspected newly arrived PR717 and PR720; their claims concern different objects.

Contributor: thakrarsagar, with substantial OpenAI Codex assistance for implementation, proof analysis, documentation and independent agent review. Apache-2.0. No pinned sprint coordination issue exists; the claim is recorded locally in CLAIM.md.

[Mister-Meeseeks]

100-PR batch review at d454ae19fdd474b4dfa7bd2c28bbe5b0996365fb.

Merged to main, preserving the original contributor commits and package attribution.

Reviewed scope: Restricted P111 mini-node factorization saves 20 native additions; complete dirty-column map and paid paths preserved, prefix changes retained explicitly.

The pinned scoped command passed; the batch ledger contains its command, input hashes, environment and log. This incorporates the research/tooling contribution and does not change the selected κ or discharge the inherited multiplication-theorem assumptions.

Batch review, attribution and evidence. Thank you for the contribution.