Content expansion, no behavior change: the four modes, the shared floor, the
redlines, and the build pipeline are untouched. This release grows the
expedition pool with six independently-verified packs and extends scene mode's
fallacy-recognition track, addressing the method-skew flagged in issue #15 (six
of the earlier packs taught a near-identical SAT-certificate lesson).
- Six new expedition packs (#15). Each was first-party verified against the
four bars (real / verified-solution / beyond-single-human / first-party-
checkable) plus a dual-use check, then double-reviewed before landing. The set
deliberately spans six distinct verification styles across six domains to
counter the earlier SAT skew:3d-euler-blowup(certified interval numerics,
fluid PDE),casp16-rna(blind-assessment negative result, RNA structure),
connect-four-bdd-oracle(symbolic BDD strong-solve, games),erdos-728-gpt5
(two-track AI-math verification),serine-hydrolase-design(forecaster,
prediction-vs-experiment de novo enzyme chemistry), and
imandra-marabou-checker(auditor, verify-the-verifier formal methods). Each
pack bakes in its honest scope so no headline overclaims survive the reveal. - Scene fallacy-recognition track grown 5→10 lenses. Added
false_analogy,
whataboutism,slippery_slope,genetic, andno_true_scotsman, each with
a defect test, reverse-guard, invariant needles, mutation coverage, and a GATE
probe. - No runtime rule, redline, or router change. Lint suite green (invariants
325/325, pack-schema 22/22, verbatim blocks 6/6), pytest 30 passed, both build
targets clean.