[Arham-Tahir64]

PR943's read-only five-source connector needs an extra pair of map calls. This contribution adds one explicitly paid operation family: temporarily XOR source rows, then restore them. For a common exact F2-linear B and arbitrary A_m, all 31 nonzero five-source output patterns use 10 B queries, 62 readouts and 12 source XORs, preserving arbitrary dirty inputs and scratch. The common-B and temporary-source-ownership assumptions are essential.

Scoped recursive-interface result; no new kappa or supplier adoption. Ten B queries are sharp when each output is read twice, even allowing arbitrarily many source shears. The proof reduces nine-query coverage to all 2,600 normalized nine-point sets in F2^5; direct pair-set and separate integer Walsh-autocorrelation checks agree on the complete histogram. Twelve XORs are an attained count, not an optimum.

The full five-factor recursive word includes every child and inverse: 2,016 arbitrary-map leaves plus 12 identity-row XORs, versus 2,072 leaves for each literal PR943 repair, with four borrowed dirty rows throughout. The invoice separately counts 648 versus 666 internal nodes and 84 versus 74 root instructions. A declared positive tariff paying leaves, XORs, nodes and root dispatch gives 2,760 versus 2,812 units. This is an abstract complete price fixture; no physical runtime advantage follows. PROOF.md states the general price inequalities and a dense implementation with separately charged access, travel, preparation and control obligations.

Only research/mutable-source-connector/ is added. Start with README.md and PROOF.md; SOURCES.json pins the three live predecessor inputs. Base/current main: bce54fb3f895cbe5cd4c0eb477c3afeed100190f; contribution: f1dcab36fa63f8226f70dd07e26f44afb9d39b51. Selector nonce a083d7387e219510c2de83d13581da77, recursive-interfaces track, refined connector-alphabet-change card. TASK.json includes the question, pilot, stop rule, local claim, exact overlapping heads and next steps.

Validation

Windows 11 / CPython 3.12.14, standard library:

python3 -B scripts/research_package.py --package research/mutable-source-connector --check-only
python3 -B scripts/research_package.py --package research/mutable-source-connector
python3 -B scripts/verify_selected_result.py

All pass, along with the card's unchanged PR943 positive control and staged whitespace checks. Evidence includes exact free noncommutative replay of all 40 incoming dirty rows in both directions, both predecessor repairs, the complete 2,600-set census, 3,024 separate dense matrix executions, nine primary mathematical controls, two concrete oracle controls and four package test groups covering optimized relocation, deterministic regeneration, occupied-output preservation and eight source/certificate corruption cases. VALIDATION.md records an initial Windows temp-directory failure and its passing scratch-directory correction.

Scope, overlap and credit

The selected record and frozen inputs are unchanged. No selected cold replay, fresh native supplier run, complete repository suite, Lean, hosted CI pass, physical supplier instantiation, ownership/frame admission, precision/tape realization or analytic-transfer proof is claimed. The next practical task is to locate an actual common-B boundary and price one source XOR completely; source immutability or an unfavorable tariff should stop that adoption attempt.

Final overlap refresh covered 821 open PRs including drafts, with complete public REST pagination at 2026-10-13 23:03:34 UTC. The repository helper searched their titles/bodies and local notes/ledger. Public discussion listing, start-here, handoff and mechanism pages were inspected; complete Discussion pagination and a connected Discussion-write tool were unavailable, so the claim is local. This is bounded overlap evidence, not proof of novelty.

Credit Ross Richardson's PR943/934, shhxera/CubeJerry's PR889, Chafik Boukhalfa's PR628, Douglas Colkitt and the retained framework. PR1185 at 4a60772285ce70f6414458f6d3a54948652bbed0 extends the unchanged alphabet to six sources; Hanson Sun's PR1215 at 8870e80d720ffa3a666bb5496c33373c45b90751 checks its free-algebra word. They are complementary and no pending code is imported. PR1103/982 borrows outputs; PR1318 concerns a different four-valued complex-phase model. No priority is claimed for classical finite-group difference bases or dirty echo cancellation.

Prepared by Arham Tahir with substantial OpenAI Codex assistance in research, algebra, both implementations, tests and writing. No subagents or independent human review. New work is Apache-2.0; predecessor sources/notices are retained unchanged.

[Mister-Meeseeks]

100-PR batch review at f1dcab36fa63f8226f70dd07e26f44afb9d39b51.

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

Reviewed scope: Mutable-source five-column connector with a sharp ten-query paired-readout bound and fully restored literal recursion. Twelve paid source shears replace two common-map calls; a favorable abstract tariff is not an established tape-runtime saving.

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.