test(major2): hver genereringstelling paret med debattens telling (SC1/SC4)

Funn 5935d942 (MISSING_TEST, tests/...:820). Briefens SC1 gjorde paringen til
vakten mot aa telle en debatt-tur som en genereringsprompt - og ingen arm i
fila asserterte debattens turantall mellom treated og control (grep «debate»
traff bare en kommentar og et filnavn).

`_debate_entries` er KOMPLEMENTET av `_GENERATION_MARK`, ikke en positiv
debatt-markoer: paastanden som gates er nettopp «ingenting som IKKE er
generering flyttet seg», og en positiv markoer ville latt en tur klassifisereren
ikke kjenner drive usett.

- T5-run: de tre genereringstellingene paret med likhet paa debatt-oppfoeringene.
- T8: fikk sink + en control-kjoering (alltid-godkjenn reviewer) - dens
  attempt-indekser [0,1] per kandidat ER en genereringstelling.
- Begge har en VAKUITETSVAKT (debatt-lista maa vaere ikke-tom): to tomme lister
  er like gratis.
- T13: record-indeksene er DOKUMENTERT som SC4s telle-proxy ved den doera -
  barnet kjoerer i egen interpreter og `--scripted-replies` har ingen
  sink-dump, saa prompt-nivaaet maales in-process (T1 for verbatim, T5-run/T8
  for paringen). Reviewens andre alternativ.

MAALT mot HELE suiten, to mutasjoner, begge roede paa NOEYAKTIG de to parede
armene og paa ingen andre: MA klassifisereren returnerer konstant tom liste
(2 roede - vakuitetsvakten) og MB filteret droppet, saa generering telles som
debatt (2 roede). Kontroll 1364 passed / 5 skipped.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

tests/test_proposal_review_loop_loadbearing.py

@@ -657,6 +657,19 @@ _VERDICT_INPUT = {"decision": "approved", "rationale": "expert reviewed (sim)"}
 #: else — the one identifier that separates a generation call from a debate turn.
 _GENERATION_MARK = "Respond with ONLY a JSON object"
 
+
+def _debate_entries(sink: list[str]) -> list[str]:
+    """Everything in the sink that is NOT a generation prompt — i.e. the debate's turns.
+
+    The complement of ``_GENERATION_MARK`` rather than a positive debate marker, because the claim
+    being guarded is exactly "nothing that is not generation moved": a positive marker would let a
+    turn the classifier does not know about drift unseen. Every "N generation prompts" assertion in
+    this file is paired with an equality of THIS list between the treated and the control run
+    (SC1/SC4) — without the pairing, an arm that counted a debate turn as a generation prompt, or a
+    door that perturbed the debate, would read as a correct generation count."""
+    return [blob for blob in sink if _GENERATION_MARK not in blob]
+
+
 # BYGG-KONTOR-NORD: affected total 300000 x 1.0 -> degenerate Monte Carlo P90 = 90000. A claim
 # <= 90000 validates; above it the deterministic validator rejects.
 _A1 = Approach(id="led-retrofit", label="Behovsstyrt belysning i fellesarealer")
@@ -755,18 +768,35 @@ async def test_t8_every_candidate_gets_its_own_keyed_answer(tmp_path: Path) -> N
     With a commissioned mandate one expert is asked N times at one terminal, so an artefact that
     cannot say WHICH candidate a sentence was about is an artefact nobody can act on. The run's
     own proposal is keyed ``OWN_PROPOSAL_ID`` — recording ``None`` there would make it
-    indistinguishable from a non-mandate run's entry, which is the property keying exists for."""
+    indistinguishable from a non-mandate run's entry, which is the property keying exists for.
+
+    The per-candidate attempt indices below are a GENERATION count (two attempts each), so SC1
+    requires them paired with the debate's count: a control run whose reviewer always approves must
+    leave the non-generation entries identical."""
+    mandate = Mandate(
+        objective="Cut energy cost without rebuilding.",
+        approaches=(_A1, _A2),
+        allow_own_proposals=True,
+    )
     reviewer = _RunReviewer()
+    treated_sink: list[str] = []
     result = await _run_with(
         outbox_dir=tmp_path / "outbox",
         reviewer=reviewer,
-        mandate=Mandate(
-            objective="Cut energy cost without rebuilding.",
-            approaches=(_A1, _A2),
-            allow_own_proposals=True,
-        ),
+        mandate=mandate,
+        sink=treated_sink,
     )
-    assert isinstance(result, RunResult)
+    control_sink: list[str] = []
+    control = await _run_with(
+        outbox_dir=tmp_path / "control",
+        run_id="run-control",
+        reviewer=_RunReviewer(revise_for=set()),
+        mandate=mandate,
+        sink=control_sink,
+    )
+    assert isinstance(result, RunResult) and isinstance(control, RunResult)
+    assert _debate_entries(treated_sink), "no debate turns reached the sink: the pairing is vacuous"
+    assert _debate_entries(treated_sink) == _debate_entries(control_sink)
     revisions = {
         (r.approach_id, r.feedback) for r in result.expert_revisions if r.decision == "revise"
     }
@@ -824,6 +854,10 @@ async def test_t5run_a_revise_on_one_approach_leaves_the_next_untouched(tmp_path
     assert len(_gen_prompts(control_sink, _A2.label)) == 1
     assert _gen_prompts(treated_sink, _A3.label) == _gen_prompts(control_sink, _A3.label)
     assert all(_FEEDBACK_SENTINEL not in b for b in _gen_prompts(treated_sink, _A3.label))
+    # SC1's pairing: the counts above mean "one more GENERATION prompt" only if the debate did not
+    # move. The non-empty check is the vacuity guard — two empty lists are equal for free.
+    assert _debate_entries(treated_sink), "no debate turns reached the sink: the pairing is vacuous"
+    assert _debate_entries(treated_sink) == _debate_entries(control_sink)
 
 
 async def test_t9_a_budget_stop_inside_generation_still_leaves_the_record(tmp_path: Path) -> None:
@@ -1129,7 +1163,16 @@ def test_t13_the_flag_answers_the_review_from_a_real_argv_and_the_answer_is_used
 
     (The plan's separate T19 — "a bare, unscripted flag parse" — is folded in here: without a
     script that argv would call a live model, so the argparse witness is this child's own
-    ``unrecognized arguments`` assertion.)"""
+    ``unrecognized arguments`` assertion.)
+
+    **The attempt indices are SC4's count PROXY at this door, and nothing more — stated rather
+    than left to be read as a prompt-level count.** The child runs in its own interpreter, so the
+    ``ScriptedChatClient`` sink lives in that process and there is no CLI flag that dumps it: the
+    prompt-level half of SC4 (the feedback VERBATIM in the second generation prompt, and the
+    debate's turn count unchanged against a control) is measured in-process, where the sink is
+    reachable — T1 for the verbatim injection, T5-run and T8 for the debate pairing. What this arm
+    settles that none of those can is that a real argv, a real pipe and a real ``run_project``
+    bought a second attempt whose OUTCOME differs (40 000, entry 2 of the step list)."""
     proc = _child(_cli_argv(tmp_path, outbox="outbox"), stdin=f"revise {_CLI_FEEDBACK}\napprove\n")
 
     assert "unrecognized arguments" not in proc.stderr, proc.stderr
@@ -1142,7 +1185,9 @@ def test_t13_the_flag_answers_the_review_from_a_real_argv_and_the_answer_is_used
         ("approve", True),
     ]
     assert payload["reviews"][0]["feedback"] == _CLI_FEEDBACK  # VERBATIM
-    assert [r["attempt"] for r in payload["reviews"]] == [0, 1]  # exactly two generation attempts
+    # SC4's count proxy at this door (see the docstring): two attempt indices, not two counted
+    # prompts — the child's sink is in the child.
+    assert [r["attempt"] for r in payload["reviews"]] == [0, 1]
 
     outcome = json.loads(
         (tmp_path / "outbox" / f"{_RUN_ID}-proposal.json").read_text(encoding="utf-8")