Pin seed->word across both languages with a shared hash

engine.ts used mulberry32 and engine.py used random.Random(seed). Same seed,
different word — so every ?seed= permalink on the site would have shown a
different puzzle than the recorded run it claimed to be replaying, and nobody
would have noticed until someone checked one by hand.

Both now derive the index from FNV-1a 32-bit over the decimal seed. A hash
rather than a PRNG because there is no honest one-line JavaScript equivalent of
Mersenne Twister, and this way there is nothing to keep in step: both sides
compute the same integer from the same string. Math.imul on the JS side is
load-bearing — a plain multiply overflows into a double and diverges after the
first few bytes.

Twelve seeds are pinned as a vector in both test suites.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019mt6sHQHEnEYrJZvoMCJSB

envs/wordle_five/tests/test_engine.py

@@ -72,6 +72,21 @@ def test_seeded_answer_is_deterministic() -> None:
     assert answer_for_seed(1) != answer_for_seed(2)
 
 
+# Pinned so the browser cannot drift. `src/demos/wordle/__tests__/engine.test.ts`
+# asserts the same twelve words for the same twelve seeds. If a language's
+# built-in RNG were used instead of the shared hash, these two lists would
+# differ and every ?seed= permalink would show a different word than the
+# recorded run it claims to replay.
+SEED_VECTORS = [
+    "wants", "amber", "spume", "toady", "divot", "filly",
+    "bobby", "clews", "hikes", "lawns", "wreak", "twist",
+]
+
+
+def test_seed_vectors_match_the_typescript_port() -> None:
+    assert [answer_for_seed(s) for s in range(12)] == SEED_VECTORS
+
+
 def test_rejections_do_not_consume_a_row() -> None:
     game = Game(seed=3, answer="tares")
     assert game.play("zzzzz")[1] is not None      # not a word