Compile native-warp frame changes with exact operator and phase checks

[Dugongue]

This adds a constructive native-warp frame-change interface: choose a new entrance U′, transport the exit by the unchanged G, explicitly emit and charge any entrance climb, and preserve every free warp operator and caller endpoint. The package includes a universal informal proof, a solver-free guarded compiler and a nontrivial complete changed BORROW-WORD. No higher κ or new supplier is claimed.

The example changes one native warp from dimension 4 to 5, edits 36 scalar frames and inserts one identity scale. Its helper inventory is unchanged; its profile changes by 28 H1 − 16 H2 − 4 H3 + 4 H4, with zero rank-mass change. This example is more expensive near the current endpoint. Its purpose is to validate the enlarged construction/search interface, not to recommend adopting a losing supplier.

Validation on Python 3.12:

• Registered package replay passes: entire compiler output equals the frozen word; exact paid frame/lifted-action and arbitrary-dirty forward, dual and cleanup-boundary checks pass.
• Five semantic rejection controls pass normally and with assertions disabled.
• The retained native receipt binds the word and the entire native source closure. A second fresh replay from the packaged native sources passed all five stages in 97.3 seconds, including independent all-column warp recipes and exact phases. NumPy is required only for this separately documented full native command.
• Source-history and commit-byte checks pass for all 44 files. The official Git-free export passes all three registered commands and all manifest hashes. No Lean or fresh kernel verification is claimed.

Reproduce with python3.12 -B scripts/research_package.py --package research/transported-warp-frames. The README separately gives the full native replay command and distinguishes retained receipts from fresh computation.

Overlap: the shared MathProof DAG already supplied common-character/hyperplane frame cuts with frozen native boundaries; this reuses that representation and constructs compatible changes across the native boundary. The opening review retrieved all 608 open PR bodies. The final main fetch remains 20ecf2891d80b5acf18d980b2164fb44ea7faa90, priority revision 6; the final updated-PR snapshot includes 36 newer bodies through #3928. Actual #3917 and #3928 sources were compared. No exhaustive novelty claim is made.

Credit: dataisfire/#3734 staged544 word; zxv/#3682 and Ross Richardson/#3605 staged predecessors; Jeremy Kothe's word/frame machinery and retained lineage; Sagar Thakrar/#3687, Arham Tahir and #3693 for the retained independent audit; Cinna Davis/#2226 and the retained exact-native lineage; Tekalign/@tekuuu for the sparse native replay. #3815's clock/packing construction remains Erwan Secousses/ChaloupeMusic's and #3875's endpoint refinement remains hexapode's. Original notices and exact native provenance are retained. New work by Dugongue with substantial OpenAI Codex assistance; no independent human review is claimed.

A favorable complete profile, chronological owner allocation, full physical caller costs and all-size/outer admission remain necessary before any κ improvement. Selected records and canonical main are unchanged; only this research package is added.