feat(types): BigInt + Decimal exact economy numbers (#52)

Types phase 2 — the "money problem". A splice-on-demand bignum engine
(runtime/native/bignum.ludic, self-contained, only intrinsics) exposed as
two namespaces:

- BigInt.* — arbitrary-precision integer (sign-magnitude, base-1e9 limbs):
  from/parse, add/sub/mul/pow, div/mod (by int), cmp/eq/is_zero, to_int, str.
  For idle counters and exact huge currencies that overflow a 32/64-bit int.
- Decimal.* — exact base-10 fixed point (BigInt mantissa + decimal scale):
  from/parse, exact add/sub/mul, cmp/eq, scale/rescale (truncate), str.
  So 0.10 + 0.20 is exactly 0.30 — no binary rounding.

Both exact => deterministic; no f32/f64. Wired: parser splice trigger
(g_uses_bignum), emit_call dispatch, reseeded seed, a self-asserting example
(examples/library/bignum.ludic + feat_case), and per-symbol docs + inventory.
All suites green incl. golden renders byte-identical and the bootstrap fixpoint.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

docs/language/bigint/_section.md

@@ -0,0 +1,7 @@
+---
+id: bigint
+title: BigInt
+order: 10
+---
+
+Arbitrary-precision integers with no upper bound, for idle/incremental counters, exact huge currencies, and score arithmetic that a 32- or 64-bit <code>int</code> would overflow. A <code>BigInt</code> is a sign-magnitude number stored as base-1e9 limbs, so every operation is <strong>exact</strong> and therefore deterministic — bit-identical on every platform, with no binary floating point. Build one with <code>BigInt.from</code> (an <code>int</code>) or <code>BigInt.parse</code> (decimal text), combine with <code>add</code> / <code>sub</code> / <code>mul</code> / <code>pow</code> / <code>div</code> / <code>mod</code>, compare with <code>cmp</code> / <code>eq</code> / <code>is_zero</code>, and render with <code>str</code>. Arguments are positional. The runtime is spliced in only when a program mentions <code>BigInt.*</code>.