[an664]

PR1368 left source 19 in R439 block F508 as an exception to its forward-label frame obstruction. This package closes that fixed eleven-wire exception: an all-length inverse-column invariant forces source 19 to contain wire 466's entry frame, which is already F508. Together with the predecessor's forward bounds, every block exit is F508 and the original {2:6, 3:2} single-climb invoice is minimal for 0<a<1, including the fixed continuation and literal inverse cleanup.

The contribution is the general forward-time inverse-column invariant and its use beyond a source-star prefix. The inverse-column support, exact data and earlier prefix obstruction remain credited to Arham Tahir's PR1368. A secondary fully charged first-contact relaxation and exact positive/failure fixtures are included.

• Coordination: claim1470, original handoff. Source head 6ae784908c57f3aa5a2165d382b21a40accdf7a6; historical main provenance 410026dc, branch directly from refreshed main ef5f3940. Fresh overlap covered 753 open PR bodies and 183 Discussion records with complete pagination and no matching new result.
• Verification: python3.12 -B scripts/research_package.py --package research/r439-source19-dual-frame-obstruction passed. Its isolated optimized replay includes the existing pinned compressed word, verified SHA256/Git blob, actual 37-ADD slice, matrix, entries, invoice and cleanup shape. No gzip is copied. Normal and optimized Python 3.12.14 runs also passed, including 13 rejection controls, 2,048 dirty inputs for a positive circuit and inverse, temporary-source-write fixtures, and 1,008 oriented GL(3,F2) updates. git diff --check passed.
• Limits: original eleven wires, exact dirty map, reversible XORs, retained monotone common-operand frames inside F508, fixed surrounding interface. No block509 result, enlarged seam, new wires, changed compiler, full 661-wire dirty replay, independent human/formal proof review, supplier/hosting admission, selected-frontier change, new kappa or full multiplication theorem. The all-length conclusion rests on the written induction, not finite tests.

Prepared by Andrey Mas with substantial GPT-6 Pro mathematical research and checker through ChatGPT; Codex coordinated, packaged and independently replayed the exact checks. Arham Tahir/OpenAI Codex and the full predecessor lineage are retained in NOTICE and source pins. Apache-2.0. Open next question: apply the same invariant to the other exceptional source under a separately pinned boundary.

[Mister-Meeseeks]

Batch review at f247c4e17b19e55c0db0ec8a1b660fd118adf014.

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

Reviewed scope: All-length forward-row/inverse-column frame induction closes source19 under monotone common-operand frames. Exact pinned eleven-wire boundary, positive fixtures and paid continuation show the original climb invoice is optimal; changed seams or nonmonotone frames are excluded.

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.