Prototype Jacurutu income, Testing Station and Shrine effects

Connect automatic unambiguous Jacurutu physical-force income, a sole-occupant
Testing Station storm choice and Shrine reciprocal physical card roles to
engine actions, player controls, legal AI and saved continuation. Preserve
Weather Control exclusion, exact card/force custody, private hand boundaries
and committed effects after restart. Shared role checks cover existing ordinary
and special Karama paths; reserved cards cannot enter another pending use.

Keep Orgiz, mixed-force Jacurutu allocation and contested Cistern/Testing
Station benefits explicitly pending. These are bounded prototypes; complete
Discovery/Advanced/expansion and publication gates remain closed.

Validation: types, lint and 4,345 offline tests passed (32 added), production
build passed, and 40 HTTP tests passed. Independent cross-reviews repaired
reserved-card deadlock, physical-card selection, alternate-role permission,
storm source/token binding and public-entitlement bot ordering. The first broad
run found five failures; focused repairs preceded the stable repeated check.

Browser A5EJQ2W4 used genuine Discovery setup/reveal/free entry with targeted QA
positioning. A human increased storm movement from two to three sectors,
moving sector 6 to 9. Refresh restored station forces and the private Snooper.
This was targeted control/recovery evidence, not an unmodified full game.
All 187 opening saved rooms remained unchanged; 31 QA/HTTP rooms were added
(218 total). Reused the running server; no restart or save reset.

Source-bound report:
/home/quintine/.local/share/dune/checkpoints/20260913-discovery-location-effects/verification-final/report.json
Source fingerprint: beecb85ab7ec39319115389c115434d161591e8549802c542a6343a0fc19bea6
Durations: check 140540ms; build 6741ms; integration 24956ms.
Aggregate goal counter 63,116,444 to 64,295,091 at final verification:
1,178,647 tokens before final Git/push overhead for three bounded functions.
Standing authorization: Push all from now on.

tests/automatic-decisions.test.ts

@@ -9,6 +9,7 @@ import {
   type Game,
 } from '../game/engine';
 import { baseDeck } from '../game/cards';
+import { placeFixtureHand } from './fixture-hand';
 const cards = baseDeck();
 const shield = cards.find((c) => c.kind === 'shield')!;
 const karama = cards.find((c) => c.effect === 'karama')!;
@@ -191,13 +192,16 @@ void test('normalization does not invent an ally split or bypass provisional unf
 void test('Truthtrance, phase opening and genuine responses retain priority over a singleton payment', () => {
   for (const overlay of ['truth', 'opening', 'response'] as const) {
     const g = pendingPayment(auction());
-    if (overlay === 'truth')
+    if (overlay === 'truth') {
+      const truth = cards.find((card) => card.effect === 'truthtrance')!;
+      placeFixtureHand(g, 2, [truth]);
       g.truthtrance = {
         stage: 'ask',
-        queue: [{ player: 'a', card: 'truth' }],
+        queue: [{ player: 'a', card: truth.id }],
         passed: [],
         question: null,
       };
+    }
     if (overlay === 'opening')
       g.phaseOpening = { initialize: false, passed: [] };
     if (overlay === 'response') {