[aliromman]

This construction raises complete conditional κ from 0.00310440546 to 0.00310626335, a gain of 0.00000185789 (0.05985%) over PR #5627. Three exact helper relations remove late scalar contacts on main fields. Each resulting H1+H5 path becomes H6. All independent inputs, productive targets and owned lifetimes stay fixed. This is a small complete gain; the larger exploratory changes have not produced another gain.

Proof · Exact relations · Complete receipt · Source pins

The complete bill has 66,492 paid calls, 12,958 productive targets and rank surplus 760: 54 fewer paid calls than the consumed word. The exact raw crossing is (0.00310626336369, 0.00310626336370), tested at 0.003106263355. Helpers, native maps, normalizers, basis work, COPY/CLEAR, routes, queries, idles and compiler completion are charged. The two interval engines, compiler and all 58 outer conditions pass. The retained ordinary BIT contract is 0.0034. The raw crossing is a ceiling only for this fixed invoice.

The exact consumed source is PR #5627 at ed4dc88a09a7d201971f6603b8c2074e95dd9c32. Its full export is retained and authenticated before imports. Task provenance is canonical main 6cf7a683f447b9f1e2333791907fd92be5009a63; the branch is based on refreshed main df0137efd35e11b4445734bf612bdd4c7ebc330c. The selected 0.00310045 record is unchanged.

Python 3.12, standard library:

python3 -B scripts/research_package.py --package research/p104-neutral-helper-cleanup

Expected: PASS_P104_NEUTRAL_HELPER_KAPPA_00310626335 and PASS_REGISTERED_RESEARCH_PACKAGE. Normal and optimized registered replays pass, including 565,504 matrix entries, 2,916 native maps and the full physical caller. Committed bytes, export integrity, read-only Git-free replay with empty PATH and six source-corruption controls pass. Validation records the boundaries. Full-history ancestry mode was not run in this shallow checkout. Hosted planning failed in the upstream Current CI navigation test: the two links for review and source_manifest are absent at the base commit. Hosted package replay was skipped. The same main regression is recorded in the workflow handoff.

Credit Ali Romman's PR #5602 boundary method and the complete PR #5627 source notices, including flyingviper's relays, Scratch-net's preparation, Pierre-Loic Doulcet's cascade and all P104/native/controller/compiler predecessors. New construction and author checks: Ali Romman with substantial OpenAI Codex assistance; no subagents. The coordination claim records the exact scope and current neighboring claims.

The five framework assumptions and helper endpoint contract remain. The huge uniform word and complete rounded multiplier are not materialized. Independent scientific acceptance and a proof-kernel theorem are not claimed. No selected-frontier file or other PR is changed.