[petalas]

Summary

ONE binary seven-space + ONE common P112 cut <=6
  -> at most THREE projective residual-rank-seven directions
  -> sharp by PR3253's existing independent triple
  -> exclude literal independent-piece caller 2^3,7^7 at code dimension 6

Add one isolated exact proof/replay package. The new universal six-cut upper bound closes a named historical caller gap; it does not recompute a current public survivor count. PR3253 owns the five-cut theorem, reconstruction libraries and six-cut lower witness. PR3289 supplies the caller context and quotient argument. Credit DreamingOfClouds/PR3154 and the retained E8/P112 geometry lineage.

The six-cut completeness check is explicit: at least 38 live root zeros still exceeds the next 37-root parabolic. All 230 affordable completion records are retained. The arbitrary-common-cut proof uses unions of actual completion cuts, with a separate 14,640-quadruple check. Six independent caller directions contradict the sharp bound three.

The standard real Weyl-chamber theorem and caller code distance 53 and capacities 2/36 remain inherited mathematical premises. The separate alpha74 seven-plus-four-flat common-two question is unchanged. No full C5, physical supplier, selected-frontier change or new kappa is claimed. Exact source pins, unchanged predecessor bytes, limits and Open next questions are included. Canonical investigation. Scientific commit 97c60446cd9103d6a8ce79134208715250db3ea9; publication 131605e644ec2aa78834d8707fe79bb05f44d169 adds only review/export receipts. All twenty reviewed package files remain unchanged.

Evidence

• Before: common-five admits at most two directions, while common-six has a known triple and no upper bound from that theorem. After: direct and dual scans agree on 230 completion records across 121 first-flat classes and prove the universal maximum three, then bind the literal caller.
• Registered CPython 3.12.15 replay, seven behavioral groups with ten actual corrupted CLI runs, Ruff E/F/B/I and formatting pass. All three actual scripts pass in the committed regular-only, inode-disjoint, read-only Git-free export under optimized CPython 3.13.16. All 22 export-file hashes stay unchanged; all 21 package/input bytes match the scientific commit.
• The designated internal read-only peer receipt is retained in REVIEW.json. This computational check is narrower than human or proof-kernel approval. Historical capacity suites, Lean, Docker and selected cold replay are unrun. Hosted CI is separate.

python3.12 -Werror -B scripts/research_package.py \
  --package research/p112-common-six-span

Merge Danger

Door: two-way

Blast Radius: isolated

One new opt-in research package. Frozen sources and selected science remain unchanged; source authentication does not establish inherited mathematics.