feat: train transformer on HF A10G + heuristic improvements

feat(data): add 20 sparse-label synthetic generators; dataset now 90 222 rows

feat: expand to 81 ciphers — improved heuristic, 12 new cipher generators, focal-loss training

chore: update cipher_examples dataset

Fix singleton class crash: normalise wallis_ciphers typo, guard train_test_split

Replace synthetic 8-label dataset with cipher-museum 100k/82-label corpus

10/10 polish: Compare Mode, Kasiski/Friedman, rich dataset schema, expanded tests/docs

Add project polish: LICENSE, CONTRIBUTING, SECURITY, CITATION, CHANGELOG, Makefile, pyproject, docs/

Initial Cipher Detective AI public foundation