[codyw912]

Change and mathematical scope

Add research/f7-l22-receipt-domain/, answering #4221 Open next question 2 without editing frozen packages. The successor checker independently reimplements #3708's necessary-condition arithmetic instead of trusting a copied eleven-split table, then binds the derived domain to four completed prior receipts plus seven deep receipts.

It checks unique split identity, completion/hits, 251-prefix coverage, recorded totals and exact source identity, with adversarial acceptance controls. No giant census, engine-soundness proof, newly established f(7)=24, new κ or selected-frontier change.

Task and overlap

Card census-pruning-audit, refined to acceptance-domain binding. Scientific base b58ad9ef4d4e861b0eaa5dba41742bdc072dca51, revision 7. Exact source pins in SOURCES.json. Nick Petalas's #4221 head 9d21222c67ecd40f49935a1e56b11f2b50d54613 and #3708 head ba1aff20ee3ce3dce35e4978feab0b6be08c7168 are credited predecessors.

Canonical main remains 624de9dc66858af0945e5f8a0942488d62280d43; guidance, selected record and scientific inputs are unchanged. Publication-time overlap refresh at 2026-10-22 22:27 UTC covered 392 open PRs, 3,875 Discussion records and all 5,862 tracked notes in the searched worktree. Exact-term matches were predecessor/own notes; neither publication branch had an existing PR in an all-state head lookup. Separate all-state title/body searches checked differently named contributions. This is not proof of novelty. Discussion post failed; no public reservation is claimed.

Validation

CPython 3.12.13, standard library:

python3.12 -B scripts/research_package.py --package research/f7-l22-receipt-domain --check-only
python3.12 -B scripts/research_package.py --package research/f7-l22-receipt-domain

Metadata and registered normal/-O runs pass: domain=11 prior=4 deep=7 controls=30 no_census_replay. The historical node sum 173,808,188,827 is authenticated recorded evidence, not freshly executed work. No C compilation, giant census, Lean build or selected construction replay. Separate read-only AI review found no blocking mathematical or acceptance defect; no independent human/formal endorsement.

Committed-byte/source-history preflight, deterministic package export integrity, and a Git-free cold export replay in normal and optimized modes also pass. The refreshed canonical launcher reproduces the same acceptance result.

Limitations

The explicit engine hash binds the seven deep rows. Prior receipt engine provenance is retained history, and three prior-complete rows are unsharded control runs rather than per-shard logs. Necessity proofs and engine soundness remain inherited.

Credit

Retains Nick Petalas, Ross Richardson, Aaron Wang, Nicholas Stone, navinor and the predecessor lineage named in README. Prepared for codyw912 with substantial OpenAI Astra assistance and separate AI review; Apache-2.0.

Open next questions

Open next questions are independently replaying the deep exclusions and formalizing the pair-identity/rotation implication, each separately scoped and budgeted. No devenv or agent configuration is included.