update

backend/tests/test_agents_api.py

@@ -0,0 +1,255 @@
+"""Direkter Text-API-Pfad (MiniMax) + RAM-Gate für opencode-Spawns."""
+
+import asyncio
+
+import httpx
+import pytest
+
+import agents
+
+TOPIC = "t"
+
+
+# ── Routing: wann läuft ein Call über die API statt über opencode? ───────────────────
+
+def test_use_text_api_routing(monkeypatch):
+    monkeypatch.setenv("MINIMAX_API_KEY", "k")
+    monkeypatch.delenv("CREATOR_TEXT_API", raising=False)
+    api = agents._use_text_api
+    assert api("minimax", "minimax/MiniMax-M3", "none", None) is True
+    assert api("minimax", "minimax-kalt/MiniMax-M2.7-highspeed", "none", None) is True
+    # Tools, Streaming, andere Provider/Modelle → Prozess-Pfad
+    assert api("minimax", "minimax/MiniMax-M3", "files", None) is False
+    assert api("minimax", "minimax/MiniMax-M3", "none", lambda s: None) is False
+    assert api("claude", "claude-sonnet-4-6", "none", None) is False
+    assert api("lokal", "ollama/qwen3.5:9b", "none", None) is False
+    # Kill-Switch und fehlender Key → Fallback
+    monkeypatch.setenv("CREATOR_TEXT_API", "0")
+    assert api("minimax", "minimax/MiniMax-M3", "none", None) is False
+    monkeypatch.delenv("CREATOR_TEXT_API")
+    monkeypatch.delenv("MINIMAX_API_KEY")
+    assert api("minimax", "minimax/MiniMax-M3", "none", None) is False
+
+
+async def test_run_agent_dispatches_to_api(monkeypatch):
+    """none+minimax → API-Runner; opencode wird nicht angefasst (auch kein which-Check)."""
+    called = {}
+
+    async def fake_api(agent_key, prompt, timeout, model, label=""):
+        called["api"] = (agent_key, model)
+        return 0, "out", "", {"input": 1, "output": 1, "reasoning": 0, "cache_read": 0, "cache_write": 0}
+
+    async def fail_oc(*a, **kw):
+        raise AssertionError("opencode-Pfad darf nicht laufen")
+
+    monkeypatch.delenv("CREATOR_FAKE_AGENTS", raising=False)
+    monkeypatch.setenv("MINIMAX_API_KEY", "k")
+    monkeypatch.setattr(agents, "_run_text_api", fake_api)
+    monkeypatch.setattr(agents, "_run_opencode", fail_oc)
+    monkeypatch.setattr(agents.shutil, "which", lambda c: None)  # API-Pfad braucht kein Binary
+    monkeypatch.setattr(agents, "resolve_role", lambda p, r: ("minimax", "minimax/MiniMax-M3"))
+    rc, out, err = await agents.run_agent("blocks-t-a", "p", 5, provider="minimax", role="judge")
+    assert (rc, out) == (0, "out") and called["api"][1] == "minimax/MiniMax-M3"
+
+
+async def test_run_agent_kill_switch_uses_opencode(monkeypatch):
+    called = {}
+
+    async def fake_oc(agent_key, prompt, timeout, provider, model, capabilities, on_line=None, label=""):
+        called["oc"] = agent_key
+        return 0, "out", ""
+
+    monkeypatch.delenv("CREATOR_FAKE_AGENTS", raising=False)
+    monkeypatch.setenv("MINIMAX_API_KEY", "k")
+    monkeypatch.setenv("CREATOR_TEXT_API", "0")
+    monkeypatch.setattr(agents, "_run_opencode", fake_oc)
+    monkeypatch.setattr(agents.shutil, "which", lambda c: "/bin/true")
+    monkeypatch.setattr(agents, "resolve_role", lambda p, r: ("minimax", "minimax/MiniMax-M3"))
+    rc, *_ = await agents.run_agent("blocks-t-b", "p", 5, provider="minimax")
+    assert rc == 0 and called["oc"] == "blocks-t-b"
+
+
+async def test_run_agent_api_tokens_in_event_meta(monkeypatch):
+    """API-Tokens landen im Event-Meta; die OpenCode-Session-DB wird NICHT konsultiert."""
+    recorded = []
+
+    async def sink(**kw):
+        recorded.append(kw)
+
+    async def fake_api(agent_key, prompt, timeout, model, label=""):
+        return 0, "out", "", {"input": 5, "output": 2, "reasoning": 0, "cache_read": 100, "cache_write": 0}
+
+    monkeypatch.delenv("CREATOR_FAKE_AGENTS", raising=False)
+    monkeypatch.setenv("MINIMAX_API_KEY", "k")
+    monkeypatch.setattr(agents, "on_event", sink)
+    monkeypatch.setattr(agents, "_run_text_api", fake_api)
+    monkeypatch.setattr(agents, "_session_tokens", lambda k: pytest.fail("Session-DB-Lookup im API-Pfad"))
+    monkeypatch.setattr(agents.shutil, "which", lambda c: "/bin/true")
+    monkeypatch.setattr(agents, "resolve_role", lambda p, r: ("minimax", "minimax/MiniMax-M3"))
+    rc, *_ = await agents.run_agent("blocks-t-tok", "p", 5, provider="minimax", scope=TOPIC)
+    assert rc == 0
+    assert recorded and recorded[0]["meta"]["tokens"] == {
+        "input": 5, "output": 2, "reasoning": 0, "cache_read": 100, "cache_write": 0}
+
+
+# ── API-Runner: Request-Bau, Antwort-Extraktion, Fehlerfälle ─────────────────────────
+
+class _FakeResp:
+    def __init__(self, status_code=200, data=None, text=""):
+        self.status_code = status_code
+        self._data = data or {}
+        self.text = text
+
+    def json(self):
+        return self._data
+
+
+def _fake_client(monkeypatch, seen, resp=None, exc=None):
+    class _Client:
+        def __init__(self, **kw):
+            seen["client_kw"] = kw
+
+        async def __aenter__(self):
+            return self
+
+        async def __aexit__(self, *a):
+            return False
+
+        async def post(self, url, json=None, headers=None):
+            seen.update(url=url, body=json, headers=headers)
+            if exc is not None:
+                raise exc
+            return resp
+
+    monkeypatch.setattr(agents.httpx, "AsyncClient", _Client)
+
+
+async def test_api_request_and_response(monkeypatch):
+    monkeypatch.setenv("MINIMAX_API_KEY", "geheim")
+    seen = {}
+    resp = _FakeResp(data={
+        "content": [{"type": "thinking", "thinking": "hm"},
+                    {"type": "text", "text": "A"}, {"type": "text", "text": "B"}],
+        "usage": {"input_tokens": 5, "output_tokens": 2,
+                  "cache_read_input_tokens": 100, "cache_creation_input_tokens": 1},
+        "stop_reason": "end_turn"})
+    _fake_client(monkeypatch, seen, resp=resp)
+    rc, out, err, tokens = await agents._run_text_api("k", "PROMPT", 5, "minimax-kalt/MiniMax-M3")
+    assert (rc, out, err) == (0, "AB", "")  # Thinking-Block übersprungen
+    assert tokens == {"input": 5, "output": 2, "reasoning": 0, "cache_read": 100, "cache_write": 1}
+    assert seen["url"] == agents._API_URL
+    assert seen["headers"]["x-api-key"] == "geheim"
+    assert seen["headers"]["anthropic-version"] == agents._API_VERSION
+    b = seen["body"]
+    assert b["model"] == "MiniMax-M3" and b["max_tokens"] == agents._API_MAX_TOKENS
+    assert b["messages"] == [{"role": "user", "content": "PROMPT"}]
+    assert b["temperature"] == 0.2 and b["thinking"] == {"type": "disabled"}  # kalt-Route
+
+
+async def test_api_model_options_per_route(monkeypatch):
+    monkeypatch.setenv("MINIMAX_API_KEY", "k")
+    seen = {}
+    resp = _FakeResp(data={"content": [{"type": "text", "text": "x"}], "usage": {}})
+    _fake_client(monkeypatch, seen, resp=resp)
+    await agents._run_text_api("k", "p", 5, "minimax-kalt/MiniMax-M2.7-highspeed")
+    assert seen["body"]["temperature"] == 0.3 and "thinking" not in seen["body"]
+    await agents._run_text_api("k", "p", 5, "minimax/MiniMax-M3")  # nativ: Endpunkt-Defaults
+    assert "temperature" not in seen["body"] and "thinking" not in seen["body"]
+
+
+async def test_api_errors(monkeypatch):
+    monkeypatch.setenv("MINIMAX_API_KEY", "k")
+    seen = {}
+    _fake_client(monkeypatch, seen, resp=_FakeResp(status_code=500, text="kaputt"))
+    rc, out, err, tokens = await agents._run_text_api("k", "p", 5, "minimax/MiniMax-M3")
+    assert rc == 1 and out == "" and "HTTP 500" in err and tokens is None
+
+    _fake_client(monkeypatch, seen, exc=httpx.ConnectError("down"))
+    rc, _, err, _ = await agents._run_text_api("k", "p", 5, "minimax/MiniMax-M3")
+    assert rc == 1 and "ConnectError" in err
+
+    _fake_client(monkeypatch, seen, exc=httpx.ReadTimeout("langsam"))
+    with pytest.raises(asyncio.TimeoutError):
+        await agents._run_text_api("k", "p", 5, "minimax/MiniMax-M3")
+
+    # leere Antwort (nur Thinking) → rc 1, Tokens bleiben sichtbar
+    resp = _FakeResp(data={"content": [{"type": "thinking", "thinking": "…"}],
+                           "usage": {"input_tokens": 3}, "stop_reason": "end_turn"})
+    _fake_client(monkeypatch, seen, resp=resp)
+    rc, _, err, tokens = await agents._run_text_api("k", "p", 5, "minimax/MiniMax-M3")
+    assert rc == 1 and "empty response" in err and tokens["input"] == 3
+
+
+async def test_api_truncation_flagged(monkeypatch):
+    monkeypatch.setenv("MINIMAX_API_KEY", "k")
+    seen = {}
+    resp = _FakeResp(data={"content": [{"type": "text", "text": "halb"}],
+                           "usage": {}, "stop_reason": "max_tokens"})
+    _fake_client(monkeypatch, seen, resp=resp)
+    rc, out, err, _ = await agents._run_text_api("k", "p", 5, "minimax/MiniMax-M3")
+    assert rc == 0 and out == "halb" and "max_tokens" in err
+
+
+# ── RAM-Gate ─────────────────────────────────────────────────────────────────────────
+
+@pytest.fixture
+def ram_gate(monkeypatch):
+    monkeypatch.setattr(agents, "RAM_MIN_FREE_PCT", 20)
+    monkeypatch.setattr(agents, "_RAM_POLL_S", 0.01)
+    monkeypatch.setattr(agents, "_opencode_recent_starts", [])
+    return monkeypatch
+
+
+async def test_ram_gate_admits_with_free_ram(ram_gate):
+    ram_gate.setattr(agents, "_opencode_running", 5)
+    ram_gate.setattr(agents, "_meminfo", lambda: (4_000_000, 8_000_000))  # 50 % frei
+    assert await agents._ram_gate("k") is True
+    assert len(agents._opencode_recent_starts) == 1  # Commit registriert
+
+
+async def test_ram_gate_waits_when_low(ram_gate):
+    ram_gate.setattr(agents, "_opencode_running", 5)
+    ram_gate.setattr(agents, "_meminfo", lambda: (800_000, 8_000_000))  # 10 % frei
+    with pytest.raises(asyncio.TimeoutError):
+        await asyncio.wait_for(agents._ram_gate("k"), 0.1)
+    # RAM wird frei → Gate lässt nach ≥1 Poll durch
+    vals = iter([(800_000, 8_000_000)])
+    ram_gate.setattr(agents, "_meminfo", lambda: next(vals, (4_000_000, 8_000_000)))
+    assert await agents._ram_gate("k") is True
+
+
+async def test_ram_gate_floor_and_fail_open(ram_gate):
+    ram_gate.setattr(agents, "_meminfo", lambda: (100_000, 8_000_000))  # fast nichts frei
+    ram_gate.setattr(agents, "_opencode_running", 1)  # unter Floor
+    assert await agents._ram_gate("k") is True
+    ram_gate.setattr(agents, "_opencode_running", 5)
+    ram_gate.setattr(agents, "_meminfo", lambda: None)  # kein /proc/meminfo
+    assert await agents._ram_gate("k") is True
+    ram_gate.setattr(agents, "RAM_MIN_FREE_PCT", 0)  # Gate aus
+    ram_gate.setattr(agents, "_meminfo", lambda: pytest.fail("Gate aus liest kein meminfo"))
+    assert await agents._ram_gate("k") is True
+
+
+async def test_ram_gate_commit_accounting(ram_gate):
+    """Knapp über der Schwelle, aber 2 frische Zulassungen → deren geschätzter RSS zählt."""
+    import time as _time
+    ram_gate.setattr(agents, "_opencode_running", 5)
+    ram_gate.setattr(agents, "_meminfo", lambda: (1_700_000, 8_000_000))  # 21 % frei
+    agents._opencode_recent_starts.extend([_time.monotonic(), _time.monotonic()])
+    with pytest.raises(asyncio.TimeoutError):
+        await asyncio.wait_for(agents._ram_gate("k"), 0.1)
+
+
+async def test_ram_gate_cancelled_scope(ram_gate):
+    ram_gate.setattr(agents, "_opencode_running", 5)
+    ram_gate.setattr(agents, "_meminfo", lambda: (800_000, 8_000_000))
+    agents.cancel_scope("blocks-cxl-")
+    try:
+        assert await agents._ram_gate("blocks-cxl-x") is False
+    finally:
+        agents.clear_scope("blocks-cxl-")
+
+
+def test_meminfo_reads_proc():
+    mem = agents._meminfo()
+    assert mem is not None and 0 < mem[0] <= mem[1]  # Linux-Testumgebung

backend/tests/test_block_calls.py

@@ -0,0 +1,538 @@
+"""Verschmolzene Board-2-Calls (block_calls.py): Generate-Konsens, Verify-Faltung mit
+Fix-Tail, Artefakte in einem Durchgang — Agenten gefaked, gegen Test-DB."""
+
+import json
+
+import pytest
+
+import block_calls as bc
+import blocks as blx
+import board_artefacts as ba
+from pipeline import FAILED, OK, GenContext
+from textkit import _norm_title
+
+TOPIC = "t"
+
+
+def _ctx():
+    return GenContext(topic=TOPIC, provider="claude", is_cancelled=lambda: False)
+
+
+def _sub(title, level="beginner", relevance="relevant", kp=None, cf=None):
+    return {"title": title, "level": level, "relevance": relevance,
+            "key_points": [f"kp {title}"] if kp is None else kp,
+            "prerequisites": "", "hurdles": "", "cited_facts": cf or [], "example_idea": ""}
+
+
+# ── Schemas ─────────────────────────────────────────────────────────────────────────
+
+def test_gen_schema_normalisiert():
+    """Gültige Einträge werden normalisiert; ungültiges level/relevance fällt auf ""
+    (Stimme entfällt, der Sub bleibt); Einträge ohne Titel fliegen."""
+    out = bc._gen_schema({"subs": [
+        {"title": " **A** ", "level": "Beginner", "relevance": "RELEVANT",
+         "key_points": ["k", ""], "cited_facts": [{"text": "t", "source": " s "},
+                                                  {"text": ""}, "quatsch"]},
+        {"title": "B", "level": "profi", "relevance": "mittel"},
+        {"title": "  "},
+    ]})
+    assert [e["title"] for e in out] == ["A", "B"]
+    assert out[0]["level"] == "beginner" and out[0]["relevance"] == "relevant"
+    assert out[0]["key_points"] == ["k"]
+    assert out[0]["cited_facts"] == [{"text": "t", "source": "s"}]
+    assert out[1]["level"] == "" and out[1]["relevance"] == ""
+
+
+def test_gen_schema_kaputt_ist_none():
+    assert bc._gen_schema(None) is None
+    assert bc._gen_schema({"subs": "x"}) is None
+    assert bc._gen_schema({"subs": []}) is None
+    assert bc._gen_schema({"subs": [{"level": "beginner"}]}) is None  # nur titellose Einträge
+
+
+def test_verify_schema_pflichtkeys_und_leeres_verdikt():
+    """Mindestens EIN Pflicht-Key muss da sein; leere Listen heißen „alles ok"."""
+    assert bc._verify_schema({}, 3) is None
+    assert bc._verify_schema({"irgendwas": 1}, 3) is None
+    v = bc._verify_schema({"gruppen": []}, 3)
+    assert v["gruppen"] == [] and v["fremd"] == set() and v["luecken"] == []
+    assert v["uebernehmen"] == {} and v["facts_probleme"] == [] and v["levels"] == {}
+
+
+def test_verify_schema_grenzen_und_normalisierung():
+    """ids außerhalb 1..n und bools fallen raus; Ein-Element-Gruppen zählen nicht;
+    uebernehmen/levels werden casefolded bzw. enum-geprüft."""
+    v = bc._verify_schema({
+        "gruppen": [{"haupt": 2, "weitere": [1, 9, True]}, {"haupt": 3, "weitere": []}],
+        "kataloge": [{"titel": " K ", "mitglieder": [1, 2]}, {"titel": "", "mitglieder": [1, 2]}],
+        "fremd": [True, 1, "2", 9],
+        "luecken": ["  x  ", "", 7],
+        "uebernehmen": {"3": " JA ", "9": "ja"},
+        "facts_probleme": [{"nr": 2, "discard": 1, "hinweis": " h "}, {"nr": 9}, "quatsch"],
+        "levels": {"1": "expert", "2": "quatsch"},
+        "relevanz": {"1": "peripheral"},
+    }, 3)
+    assert v["gruppen"] == [{"haupt": 2, "ids": [1, 2]}]
+    assert v["kataloge"] == [{"titel": "K", "ids": [1, 2]}]
+    assert v["fremd"] == {1, 2}
+    assert v["luecken"] == ["x"]
+    assert v["uebernehmen"] == {3: "ja"}
+    assert v["facts_probleme"] == [{"nr": 2, "discard": True, "hinweis": "h"}]
+    assert v["levels"] == {1: "expert"} and v["relevanz"] == {1: "peripheral"}
+
+
+def test_art_gen_schema_pattern_ist_pflicht():
+    """Ohne verwertbares pattern kein Verdikt (Leitner hängt an den Fragen);
+    cards/examples sind best-effort und werden einzeln validiert."""
+    assert bc._art_gen_schema({"cards": [], "examples": []}) is None
+    assert bc._art_gen_schema("x") is None
+    out = bc._art_gen_schema({
+        "pattern": [{"block": "B", "subblock": "S", "question": "F?"},
+                    {"block": "B", "subblock": "", "question": "F?"}],
+        "cards": [{"block": "B", "subblock": "S", "question": "F?", "answer": "A"},
+                  {"block": "B", "subblock": "S", "question": "F?"}],
+        "examples": [{"block": "B", "subblock": "S", "problem": "P", "steps": ["s1", ""], "result": ""},
+                     {"block": "B", "subblock": "S", "problem": "P", "steps": []}],
+    })
+    assert len(out["pattern"]) == 1 and len(out["cards"]) == 1
+    assert out["examples"] == [{"block": "B", "subblock": "S", "problem": "P",
+                                "steps": ["s1"], "result": ""}]
+
+
+def test_art_check_schema_varianten():
+    """{"ok": true} → leeres Verdikt; ohne bekannten Key None; Beispiel-Indizes sind
+    1-basiert, bools/0 zählen nicht."""
+    ok = bc._art_check_schema({"ok": True})
+    assert ok == {"pattern": [], "pattern_ergaenzt": [], "examples_probleme": set()}
+    assert bc._art_check_schema({"foo": 1}) is None
+    v = bc._art_check_schema({"examples_probleme": [1, "2", {"index": 3}, True, 0, -1],
+                              "pattern_ergaenzt": [{"block": "B", "subblock": "S", "question": "F?"}]})
+    assert v["examples_probleme"] == {1, 2, 3}
+    assert len(v["pattern_ergaenzt"]) == 1 and v["pattern"] == []
+
+
+# ── Generate ────────────────────────────────────────────────────────────────────────
+
+@pytest.fixture
+def env(testdb, tmp_path, monkeypatch):
+    """Ohne Modell (exakte Norm-Gleichheit), ohne Korpus (thema-Selbst-Recherche)."""
+    monkeypatch.setattr(bc, "EMBEDDING_AKTIV", False)
+    monkeypatch.setattr(blx, "EMBEDDING_AKTIV", False)  # _dedup_subblocks aus
+    monkeypatch.setattr(bc, "material_folder", lambda t: None)
+    monkeypatch.setattr(bc, "load_source", lambda t: {"type": "thema"})
+    return testdb, _ctx(), {"arbeit": tmp_path}
+
+
+def _mk_gen_race(outputs):
+    """_race-Fake: pro Generator-Slot (…-gN) die gescriptete Antwort als Text;
+    fehlender Eintrag = Ausfall."""
+    calls = []
+
+    async def fake_race(topic, label, slots, quorum, timeout, provider, on_update=None,
+                        cancelled=None, **kw):
+        outs = []
+        for slot in slots:
+            calls.append(slot["key"])
+            g = int(slot["key"].rsplit("-g", 1)[1])
+            out = outputs.get(g)
+            if out is not None:
+                outs.append(slot["payload"]((0, json.dumps(out), "")))
+        return [o for o in outs if o] or None
+
+    fake_race.calls = calls
+    return fake_race
+
+
+async def test_generate_schnittmenge_wird_consensus(env, monkeypatch):
+    """Von beiden Generatoren genannt → consensus (Facts-Union); Einzelnennungen
+    werden unsicher und gehen zum Prüfer."""
+    db, ctx, files = env
+    monkeypatch.setattr(bc, "_race", _mk_gen_race({
+        1: {"subs": [_sub("Sub A", kp=["k1"]), _sub("Sub B")]},
+        2: {"subs": [_sub("Sub A", kp=["k2"]), _sub("Sub C")]},
+    }))
+    gen = await bc._generate_block(ctx, files, "Alpha", "Grundkonzept")
+    assert gen["raw"] == {"Alpha": ["Sub A"]}
+    assert gen["facts"]["Alpha"]["sub a"]["key_points"] == ["k1", "k2"]  # Union beider Nennungen
+    assert {u["title"] for u in gen["unsicher"]} == {"Sub B", "Sub C"}
+    assert gen["votes"]["sub a"]["level"] == ["beginner", "beginner"]
+    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "alpha")}
+    assert rows["sub a"] == "consensus"
+    assert rows["sub b"] == rows["sub c"] == "candidate"
+
+
+async def test_generate_degraded_alles_unsicher(env, monkeypatch):
+    """Liefert nur EIN Generator, ist kein Konsens möglich — alles wird unsicher,
+    der Prüfer entscheidet mit Material."""
+    db, ctx, files = env
+    monkeypatch.setattr(bc, "_race", _mk_gen_race({
+        1: {"subs": [_sub("Sub A"), _sub("Sub B")]},  # g2 fällt aus
+    }))
+    gen = await bc._generate_block(ctx, files, "Alpha", "Grundkonzept")
+    assert gen["raw"] == {"Alpha": []}
+    assert {u["title"] for u in gen["unsicher"]} == {"Sub A", "Sub B"}
+    assert not any(r["status"] == "consensus" for r in await db.list_subblocks(TOPIC, "alpha"))
+
+
+async def test_generate_beide_ausgefallen_ist_none(env, monkeypatch):
+    db, ctx, files = env
+    monkeypatch.setattr(bc, "_race", _mk_gen_race({}))
+    assert await bc._generate_block(ctx, files, "Alpha", "d") is None
+
+
+async def test_generate_seed_garantie(env, monkeypatch):
+    """Ungedeckte Seeds gehen als unsicher zum Prüfer (Beleg-Gate liegt dort);
+    lexikalisch gedeckte Seeds erzeugen keine Dublette."""
+    db, ctx, files = env
+    monkeypatch.setattr(bc, "_race", _mk_gen_race({
+        1: {"subs": [_sub("Sub A")]}, 2: {"subs": [_sub("Sub A")]},
+    }))
+    gen = await bc._generate_block(ctx, files, "Alpha", "d",
+                                   seeds=["Escaping Regeln", "Sub"])
+    assert gen["raw"] == {"Alpha": ["Sub A"]}
+    assert [u["title"] for u in gen["unsicher"]] == ["Escaping Regeln"]  # „Sub" ist gedeckt
+    assert gen["unsicher"][0]["key_points"] == []  # Seeds kommen ohne Beleg
+
+
+async def test_generate_resume_ohne_neue_calls(env, monkeypatch):
+    """Vorhandene gen-Dateien → kein neuer _race-Call, Ergebnis wird übernommen."""
+    db, ctx, files = env
+    fake = _mk_gen_race({1: {"subs": [_sub("Sub A")]}, 2: {"subs": [_sub("Sub A")]}})
+    monkeypatch.setattr(bc, "_race", fake)
+    gen1 = await bc._generate_block(ctx, files, "Alpha", "d")
+    n = len(fake.calls)
+    gen2 = await bc._generate_block(ctx, files, "Alpha", "d")
+    assert len(fake.calls) == n  # alles resumed
+    assert gen2["raw"] == gen1["raw"]
+
+
+# ── Verify (+ Fix-Tail) ─────────────────────────────────────────────────────────────
+
+def _gen_von(title, subs, unsicher=None, votes=None):
+    """Karten-Payload wie aus _generate_block: raw/facts/unsicher/votes."""
+    return {"raw": {title: list(subs)},
+            "facts": {title: {_norm_title(s): {"key_points": [f"kp {s}"], "prerequisites": "",
+                                               "hurdles": "", "cited_facts": [], "example_idea": ""}
+                              for s in subs}},
+            "unsicher": unsicher or [], "votes": votes or {}}
+
+
+def _judge_slot(antworten, fix=None):
+    """run_single_slot-Fake: Prüfer-Antwort je j-Suffix, Fix-Antwort für -sb-fix-;
+    fehlender Eintrag = FAILED."""
+    calls = []
+
+    async def fake(ctx, label, *, key, prompt, role, capabilities, payload, timeout, on_line=None):
+        calls.append({"key": key, "prompt": prompt})
+        if "-sb-fix-" in key:
+            if fix is None:
+                return FAILED, None
+            return OK, payload((0, json.dumps(fix), ""))
+        j = key.rsplit("-j", 1)[-1]
+        antwort = antworten.get(j)
+        if antwort is None:
+            return FAILED, None
+        return OK, payload((0, json.dumps(antwort), ""))
+
+    fake.calls = calls
+    return fake
+
+
+async def _seed_rows(db, bnorm, titles, status="consensus"):
+    for t in titles:
+        await db.put_subblock(TOPIC, bnorm, _norm_title(t), bnorm.title(), t, status=status)
+
+
+async def test_verify_gruppe_faltet_einstimmig(env, monkeypatch):
+    """Beide Prüfer gruppieren 1+2 → haupt gewinnt, Verlierer wird variant und seine
+    Facts wandern per Union zum Gewinner; Dissens-Gruppen falten nicht."""
+    db, ctx, files = env
+    subs = ["Marker Regel", "Marker Regel im Detail erklärt", "Eigenes Thema"]
+    await _seed_rows(db, "alpha", subs)
+    verdikt = {"gruppen": [{"haupt": 2, "weitere": [1]}]}
+    fake = _judge_slot({"1": verdikt, "2": {"gruppen": [{"haupt": 2, "weitere": [1]},
+                                                        {"haupt": 3, "weitere": [2]}]}})
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs), {})
+    assert res["raw"] == {"Alpha": [subs[1], subs[2]]}  # Gruppe 2+3 war einseitig → kein Fold
+    wf = res["facts"]["Alpha"][_norm_title(subs[1])]
+    assert wf["key_points"] == [f"kp {subs[1]}", f"kp {subs[0]}"]  # Union geerbt
+    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "alpha")}
+    assert rows[_norm_title(subs[0])] == "variant"
+    assert rows[_norm_title(subs[1])] == "consensus"
+
+
+async def test_verify_fremd_nur_einstimmig(env, monkeypatch):
+    """Fremd 2/2 → discarded + raus; einseitig fremd → bleibt."""
+    db, ctx, files = env
+    subs = ["CSS Regel", "Echte Regel"]
+    await _seed_rows(db, "alpha", subs)
+    fake = _judge_slot({"1": {"fremd": [1, 2]}, "2": {"fremd": [1]}})
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs), {})
+    assert res["raw"] == {"Alpha": ["Echte Regel"]}
+    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "alpha")}
+    assert rows["css regel"] == "discarded" and rows["echte regel"] == "consensus"
+
+
+async def test_verify_uebernahme_braucht_beide(env, monkeypatch):
+    """Unsicher-Eintrag wird nur bei 2/2 „ja" consensus (samt Generator-Facts);
+    sonst discarded."""
+    db, ctx, files = env
+    subs = ["Sub A"]
+    unsicher = [_sub("Unsicher B", kp=["kp b"]), _sub("Unsicher C")]
+    await _seed_rows(db, "alpha", subs)
+    await _seed_rows(db, "alpha", ["Unsicher B", "Unsicher C"], status="candidate")
+    fake = _judge_slot({"1": {"uebernehmen": {"2": "ja", "3": "ja"}},
+                        "2": {"uebernehmen": {"2": "ja", "3": "nein"}}})
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs, unsicher=unsicher), {})
+    assert res["raw"] == {"Alpha": ["Sub A", "Unsicher B"]}
+    assert res["facts"]["Alpha"]["unsicher b"]["key_points"] == ["kp b"]
+    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "alpha")}
+    assert rows["unsicher b"] == "consensus" and rows["unsicher c"] == "discarded"
+    # der Prüfer-Prompt weist die Unsicher-Nummern aus
+    assert "UNSICHER" in fake.calls[0]["prompt"] and "entries 2–3" in fake.calls[0]["prompt"]
+
+
+async def test_verify_facts_discard_nur_2von2(env, monkeypatch):
+    """Facts-Discard ist irreversibel → nur 2/2; die einseitige Stimme ohne Hinweis
+    löst auch keinen Fix aus."""
+    db, ctx, files = env
+    subs = ["Sub A", "Sub B"]
+    await _seed_rows(db, "alpha", subs)
+    fake = _judge_slot({"1": {"facts_probleme": [{"nr": 1, "discard": True},
+                                                 {"nr": 2, "discard": True}]},
+                        "2": {"facts_probleme": [{"nr": 1, "discard": True}]}})
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs), {})
+    assert res["raw"] == {"Alpha": ["Sub B"]}
+    assert not any("-sb-fix-" in c["key"] for c in fake.calls)
+    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "alpha")}
+    assert rows["sub a"] == "discarded" and rows["sub b"] == "consensus"
+
+
+async def test_verify_korrektur_ab_einer_stimme(env, monkeypatch):
+    """Ein Hinweis EINES Prüfers reicht: der Fix-Call läuft und ersetzt die Facts des
+    beanstandeten Subs; die Einstufung bleibt."""
+    db, ctx, files = env
+    subs = ["Sub A", "Sub B"]
+    await _seed_rows(db, "alpha", subs)
+    fix = {"subs": [_sub("Sub A", kp=["korrigierte Aussage"])]}
+    fake = _judge_slot({"1": {"facts_probleme": [{"nr": 1, "hinweis": "Zahl falsch"}]},
+                        "2": {"gruppen": []}}, fix=fix)
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs), {})
+    assert any("-sb-fix-" in c["key"] for c in fake.calls)
+    side = {s["title"]: s for s in res["sidecar"]["Alpha"]}
+    assert side["Sub A"]["facts"]["key_points"] == ["korrigierte Aussage"]
+    assert res["facts"]["Alpha"]["sub a"]["key_points"] == ["korrigierte Aussage"]
+    assert side["Sub B"]["facts"]["key_points"] == ["kp Sub B"]  # unbeanstandet
+
+
+async def test_verify_luecke_belegt_wird_neuer_sub(env, monkeypatch):
+    """Lücken-Schnitt beider Prüfer → Fix legt den belegten Fund als neuen consensus-Sub
+    an; ein unbelegter „Fund" verfällt am Beleg-Gate."""
+    db, ctx, files = env
+    subs = ["Sub A"]
+    await _seed_rows(db, "alpha", subs)
+    fix = {"subs": [_sub("Escaping von Sonderzeichen", level="expert", kp=["belegt"]),
+                    _sub("Unbelegte Behauptung", kp=[])]}
+    fake = _judge_slot({"1": {"luecken": ["Escaping fehlt"]},
+                        "2": {"luecken": ["Escaping unbehandelt"]}}, fix=fix)
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs), {})
+    assert res["raw"] == {"Alpha": ["Sub A", "Escaping von Sonderzeichen"]}
+    neu = next(s for s in res["sidecar"]["Alpha"] if s["title"] == "Escaping von Sonderzeichen")
+    assert neu["level"] == "expert" and neu["facts"]["key_points"] == ["belegt"]
+    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "alpha")}
+    assert rows[_norm_title("Escaping von Sonderzeichen")] == "consensus"
+    assert _norm_title("Unbelegte Behauptung") not in rows
+
+
+async def test_verify_ersatzrichter_bei_ausfall(env, monkeypatch):
+    """Fällt EIN Prüfer aus, springt der Ersatz jE ein — Einstimmigkeit mit ihm faltet."""
+    db, ctx, files = env
+    subs = ["CSS Regel", "Echte Regel"]
+    await _seed_rows(db, "alpha", subs)
+    fake = _judge_slot({"1": {"fremd": [1]}, "E": {"fremd": [1]}})  # j2 → FAILED
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs), {})
+    assert res["raw"] == {"Alpha": ["Echte Regel"]}
+    assert [c["key"].rsplit("-j", 1)[-1] for c in fake.calls] == ["1", "2", "E"]
+
+
+async def test_verify_failopen_verwirft_nur_unsicher(env, monkeypatch):
+    """Nur 1 Prüfer (auch der Ersatz fällt aus) → fail-open: consensus bleibt unangetastet,
+    unsicher wird verworfen (ohne Panel keine Übernahme-Entscheidung)."""
+    db, ctx, files = env
+    subs = ["Sub A"]
+    unsicher = [_sub("Unsicher B")]
+    await _seed_rows(db, "alpha", subs)
+    await _seed_rows(db, "alpha", ["Unsicher B"], status="candidate")
+    fake = _judge_slot({"1": {"fremd": [1], "uebernehmen": {"2": "ja"}}})  # j2+jE → FAILED
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs, unsicher=unsicher), {})
+    assert res["raw"] == {"Alpha": ["Sub A"]}  # fremd-Einzelstimme wirkt NICHT
+    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "alpha")}
+    assert rows["sub a"] == "consensus" and rows["unsicher b"] == "discarded"
+
+
+async def test_verify_level_korrektur_wiegt_doppelt(env, monkeypatch):
+    """Explizite Prüfer-Korrektur (×2) schlägt die Generator-Stimme; Patt fällt auf
+    advanced/relevant (heutige Defaults)."""
+    db, ctx, files = env
+    subs = ["Sub A", "Sub B"]
+    await _seed_rows(db, "alpha", subs)
+    votes = {"sub a": {"level": ["beginner"], "relevance": []},
+             "sub b": {"level": ["beginner", "expert"], "relevance": []}}
+    fake = _judge_slot({"1": {"levels": {"1": "expert"}}, "2": {"gruppen": []}})
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs, votes=votes), {})
+    side = {s["title"]: s for s in res["sidecar"]["Alpha"]}
+    assert side["Sub A"]["level"] == "expert"     # 2× Korrektur > 1× Generator
+    assert side["Sub B"]["level"] == "advanced"   # 1:1-Patt → Default
+    assert side["Sub A"]["relevance"] == "relevant"  # keine Stimme → Default
+
+
+async def test_verify_resume_ohne_neue_calls(env, monkeypatch):
+    """Vorhandene verify-j-Dateien → kein neuer Prüfer-Call."""
+    db, ctx, files = env
+    subs = ["Sub A"]
+    await _seed_rows(db, "alpha", subs)
+    fake = _judge_slot({"1": {"gruppen": []}, "2": {"gruppen": []}})
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs), {})
+    n = len(fake.calls)
+    await bc._verify_block(ctx, files, "Alpha", _gen_von("Alpha", subs), {})
+    assert len(fake.calls) == n
+
+
+# ── Artefakte ───────────────────────────────────────────────────────────────────────
+
+def _sidecar(titles):
+    return [{"title": t, "level": "beginner", "relevance": "relevant",
+             "facts": {"key_points": [f"kp {t}"]}} for t in titles]
+
+
+def _art_slot(gen_out, check_out):
+    """run_single_slot-Fake für Artefakte: gen_out je Teil (dict oder callable(prompt)),
+    check_out fürs Prüfer-Verdikt."""
+    calls = []
+
+    async def fake(ctx, label, *, key, prompt, role, capabilities, payload, timeout, on_line=None):
+        calls.append({"key": key, "prompt": prompt})
+        if "-art-gen-" in key:
+            out = gen_out(prompt) if callable(gen_out) else gen_out
+            return OK, payload((0, json.dumps(out), ""))
+        if "-art-check-" in key:
+            if check_out is None:
+                return FAILED, None
+            return OK, payload((0, json.dumps(check_out), ""))
+        raise AssertionError(f"unerwarteter Call {key}")
+
+    fake.calls = calls
+    return fake
+
+
+async def test_artefakte_ein_call_liefert_alles(env, monkeypatch):
+    """EIN Generator-Call liefert pattern+cards+examples, der Prüfer sagt ok →
+    Rohfassung wird übernommen, block-Feld auf den Karten-Block normiert."""
+    db, ctx, files = env
+    gen_out = {"pattern": [{"block": "Echo", "subblock": "Sub A", "question": "F?"}],
+               "cards": [{"block": "Echo", "subblock": "Sub A", "question": "F?", "answer": "A"}],
+               "examples": [{"block": "Echo", "subblock": "Sub A", "problem": "P",
+                             "steps": ["s1"], "result": "R"}]}
+    fake = _art_slot(gen_out, {"ok": True})
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._artefakte_block(ctx, files, "Alpha", _sidecar(["Sub A"]))
+    assert [c["key"] for c in fake.calls if "-art-gen-" in c["key"]].__len__() == 1
+    assert res["pattern"] == {"Alpha": [{"subblock": "Sub A", "question": "F?"}]}
+    assert res["artefacts"]["flashcard"] == [{"block": "Alpha", "subblock": "Sub A",
+                                              "question": "F?", "answer": "A"}]
+    assert res["artefacts"]["example"][0]["block"] == "Alpha"  # Agent-Echo „Echo" normiert
+
+
+async def test_artefakte_check_entfernt_beispiel_und_ergaenzt_frage(env, monkeypatch):
+    """examples_probleme wirft das beanstandete Beispiel; pattern_ergaenzt füllt die
+    fehlende Frage nach — der Prüfer-Prompt listet den fraglosen Sub."""
+    db, ctx, files = env
+    gen_out = {"pattern": [{"block": "Alpha", "subblock": "Sub A", "question": "F?"}],
+               "cards": [],
+               "examples": [{"block": "Alpha", "subblock": "Sub A", "problem": "P1", "steps": ["x"], "result": ""},
+                            {"block": "Alpha", "subblock": "Sub A", "problem": "P2", "steps": ["y"], "result": ""}]}
+    check = {"examples_probleme": [{"index": 1}],
+             "pattern_ergaenzt": [{"block": "Alpha", "subblock": "Sub B", "question": "F B?"}]}
+    fake = _art_slot(gen_out, check)
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._artefakte_block(ctx, files, "Alpha", _sidecar(["Sub A", "Sub B"]))
+    assert [e["problem"] for e in res["artefacts"]["example"]] == ["P2"]
+    assert res["pattern"]["Alpha"] == [{"subblock": "Sub A", "question": "F?"},
+                                       {"subblock": "Sub B", "question": "F B?"}]
+    check_prompt = next(c["prompt"] for c in fake.calls if "-art-check-" in c["key"])
+    assert "STILL MISSING A QUESTION" in check_prompt and "Sub B" in check_prompt
+
+
+async def test_artefakte_split_ab_schwelle(env, monkeypatch):
+    """> ART_SPLIT_SUBS Subs → ZWEI parallele Generator-Calls, jeder sieht nur seine
+    Hälfte; die Ergebnisse werden zusammengeführt."""
+    db, ctx, files = env
+    monkeypatch.setattr(bc, "ART_SPLIT_SUBS", 2)
+
+    def gen_out(prompt):
+        subs = [t for t in ("Sub A", "Sub B", "Sub C") if f"- {t}" in prompt]
+        return {"pattern": [{"block": "Alpha", "subblock": s, "question": f"F {s}?"} for s in subs],
+                "cards": [], "examples": []}
+
+    fake = _art_slot(gen_out, {"ok": True})
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._artefakte_block(ctx, files, "Alpha", _sidecar(["Sub A", "Sub B", "Sub C"]))
+    gen_keys = [c["key"] for c in fake.calls if "-art-gen-" in c["key"]]
+    assert len(gen_keys) == 2 and gen_keys[0].endswith("-t1") and gen_keys[1].endswith("-t2")
+    assert [p["subblock"] for p in res["pattern"]["Alpha"]] == ["Sub A", "Sub B", "Sub C"]
+
+
+async def test_artefakte_check_ausfall_uebernimmt_rohfassung(env, monkeypatch):
+    """Prüfer ohne Ergebnis → fail-open, die Generator-Rohfassung zählt."""
+    db, ctx, files = env
+    gen_out = {"pattern": [{"block": "Alpha", "subblock": "Sub A", "question": "F?"}],
+               "cards": [], "examples": []}
+    fake = _art_slot(gen_out, None)
+    monkeypatch.setattr(bc, "run_single_slot", fake)
+    res = await bc._artefakte_block(ctx, files, "Alpha", _sidecar(["Sub A"]))
+    assert res["pattern"] == {"Alpha": [{"subblock": "Sub A", "question": "F?"}]}
+
+
+async def test_artefakte_leerer_block(env):
+    db, ctx, files = env
+    res = await bc._artefakte_block(ctx, files, "Alpha", [])
+    assert res == {"pattern": {"Alpha": []}, "artefacts": {"flashcard": [], "example": []}}
+
+
+# ── Migration der alten Stage-Treppe ────────────────────────────────────────────────
+
+async def test_migriere_alt_karten(testdb):
+    """Karten in alten Stages gehen mit reduziertem Payload zurück nach generate
+    (alte Zwischenstände sind für die verschmolzenen Calls wertlos); Terminal- und
+    Neu-Struktur-Karten bleiben unangetastet."""
+    db = testdb
+    await db.kanban_upsert_card(TOPIC, "artefacts", "alpha", "ablock", "facts",
+                                {"title": "Alpha", "description": "d", "n_size": 3,
+                                 "sources": ["s1"], "raw": {"Alpha": ["alt"]},
+                                 "facts": {"Alpha": {}}})
+    await db.kanban_upsert_card(TOPIC, "artefacts", "beta", "ablock", "question_pattern",
+                                {"title": "Beta", "sidecar": {}})
+    await db.kanban_upsert_card(TOPIC, "artefacts", "gamma", "ablock", "done_artefact",
+                                {"title": "Gamma", "raw": {"Gamma": ["bleibt"]}})
+    await db.kanban_upsert_card(TOPIC, "artefacts", "delta", "ablock", "verify",
+                                {"title": "Delta", "raw": {"Delta": ["neu"]}})
+    n = await ba.migriere_alt_karten(TOPIC)
+    assert n == 2
+    alpha = await db.kanban_get_card(TOPIC, "artefacts", "alpha")
+    assert alpha["stage"] == "generate"
+    assert alpha["payload"] == {"title": "Alpha", "description": "d", "n_size": 3, "sources": ["s1"]}
+    assert (await db.kanban_get_card(TOPIC, "artefacts", "beta"))["stage"] == "generate"
+    gamma = await db.kanban_get_card(TOPIC, "artefacts", "gamma")
+    assert gamma["stage"] == "done_artefact" and gamma["payload"]["raw"] == {"Gamma": ["bleibt"]}
+    delta = await db.kanban_get_card(TOPIC, "artefacts", "delta")
+    assert delta["stage"] == "verify" and delta["payload"]["raw"] == {"Delta": ["neu"]}

backend/tests/test_board_inventory.py

@@ -70,39 +70,36 @@ async def board_env(testdb, tmp_path, monkeypatch):
         return False
     monkeypatch.setattr(bi, "_emb_ok", no_emb)
 
-    async def fake_subblocks(ctx, set_p, files, entries, instructions, wipe=True, ns="", seeds=None, lbl="", sources=None):
-        title = list(entries.values())[0].split(" — ")[0]
-        return {title: ["Sub Eins", "Sub Zwei"]}
+    async def fake_generate(ctx, files, title, description, instructions="", ns="", lbl="",
+                            sources=None, seeds=None):
+        subs = ["Sub Eins", "Sub Zwei"]
+        return {"raw": {title: list(subs)},
+                "facts": {title: {_norm_title(s): {"key_points": [f"Fakt zu {s}"], "cited_facts": []}
+                                  for s in subs}},
+                "unsicher": [], "votes": {}}
 
-    async def fake_facts(ctx, set_p, files, raw, q, folder, instructions, ns="", lbl="", sources=None):
-        facts = {t: {_norm_title(s): {"key_points": [f"Fakt zu {s}"], "cited_facts": []}
-                     for s in subs} for t, subs in raw.items()}
-        return facts, {}
+    async def fake_verify(ctx, files, title, gen, q, instructions="", ns="", lbl="", sources=None):
+        subs = gen["raw"].get(title) or []
+        bfacts = gen["facts"].get(title) or {}
+        sidecar = [{"title": s, "level": "beginner",
+                    "relevance": "relevant" if i == 0 else "peripheral",
+                    "facts": bfacts.get(_norm_title(s)) or {}}
+                   for i, s in enumerate(subs)]
+        return {"raw": {title: list(subs)}, "facts": {title: bfacts}, "sidecar": {title: sidecar}}
 
-    async def fake_levels(ctx, set_p, files, raw, instructions, ns="", lbl=""):
-        return {t: [{"title": s, "level": "beginner"} for s in subs] for t, subs in raw.items()}
-
-    async def fake_relevance(ctx, set_p, files, sidecar, instructions, ns="", lbl=""):
-        return {1: "relevant", 2: "peripheral"}
-
-    async def fake_pattern(ctx, set_p, files, sidecar, instructions, ns="", lbl=""):
-        return {t: [{"subblock": subs[0]["title"], "question": f"Was ist {t}?"}]
-                for t, subs in sidecar.items()}
-
-    async def fake_artefacts(ctx, set_p, files, sidecar, instructions, ns="", lbl=""):
-        return {"flashcard": [{"block": t, "subblock": subs[0]["title"], "front": "F", "back": "B"}
-                              for t, subs in sidecar.items()], "example": []}
+    async def fake_artefakte(ctx, files, title, sidecar_subs, instructions="", ns="", lbl=""):
+        if not sidecar_subs:
+            return {"pattern": {title: []}, "artefacts": {"flashcard": [], "example": []}}
+        first = sidecar_subs[0]["title"]
+        return {"pattern": {title: [{"subblock": first, "question": f"Was ist {title}?"}]},
+                "artefacts": {"flashcard": [{"block": title, "subblock": first, "front": "F", "back": "B"}],
+                              "example": []}}
 
     async def fake_outline(ctx, set_p, files, entries, instructions):
         return {"chapters": [{"title": "Kapitel 1", "numbers": sorted(entries)}]}
 
-    async def fake_konsolidierung(ctx, files, raw, facts_map, instructions="", ns="", lbl=""):
-        return None
-
-    for name, fn in [("_subblocks_block", fake_subblocks), ("_facts_block", fake_facts),
-                     ("_levels_block", fake_levels), ("_relevance_block", fake_relevance),
-                     ("_question_pattern_block", fake_pattern), ("_artefacts_block", fake_artefacts),
-                     ("_outline_block", fake_outline), ("_konsolidiere_subblocks", fake_konsolidierung)]:
+    for name, fn in [("_generate_block", fake_generate), ("_verify_block", fake_verify),
+                     ("_artefakte_block", fake_artefakte), ("_outline_block", fake_outline)]:
         monkeypatch.setattr(ba, name, fn)
 
     class _EmbOff:  # Cross-Block-Barrier reicht ohne Modell alle Karten durch
@@ -238,16 +235,17 @@ async def test_filter_judges_run_parallel(board_env, monkeypatch):
 
 
 async def test_empty_subblocks_completes_without_deadletter(board_env, monkeypatch):
-    """Legitim leere Subbausteine ({} statt None) → Karte läuft bis done_artefact durch."""
+    """Legitim leere Subbausteine (leere raw-Liste statt None) → Karte läuft bis done_artefact durch."""
     import asyncio
     import board_artefacts as ba
     import blocks as blx
     db, ctx, files = board_env
 
-    async def empty_subs(ctx, set_p, files, entries, instructions, wipe=True, ns="", seeds=None, lbl="", sources=None):
-        return {}
-    monkeypatch.setattr(ba, "_subblocks_block", empty_subs)
-    await db.kanban_upsert_card(TOPIC, "artefacts", "leer", "ablock", "subblocks",
+    async def empty_gen(ctx, files, title, description, instructions="", ns="", lbl="",
+                        sources=None, seeds=None):
+        return {"raw": {title: []}, "facts": {title: {}}, "unsicher": [], "votes": {}}
+    monkeypatch.setattr(ba, "_generate_block", empty_gen)
+    await db.kanban_upsert_card(TOPIC, "artefacts", "leer", "ablock", "generate",
                                 {"title": "Leerer Block", "description": "d"})
     ok = await asyncio.wait_for(
         bi.run_boards(ctx, lambda *a, **k: None, files, {"type": "thema"}, None, "", research=False),
@@ -712,18 +710,18 @@ def test_agent_priority_order():
 
 
 async def test_outline_runs_before_artefacts_finish(board_env, monkeypatch):
-    """Gliederung startet, sobald alle Karten die facts-Stage passiert haben —
+    """Gliederung startet, sobald alle Karten die generate-Stage passiert haben —
     parallel zu den restlichen Artefakt-Stages des langsamsten Blocks."""
     import asyncio
     import board_artefacts as ba
     db, ctx, files = board_env
     await _seed(db)
-    base_levels = ba._levels_block
+    base_verify = ba._verify_block
     snapshot = {}
 
-    async def slow_levels(ctx, set_p, files, raw, instructions, ns="", lbl=""):
-        await asyncio.sleep(0.8)  # keeps one card in `levels` while the outline fires
-        return await base_levels(ctx, set_p, files, raw, instructions, ns=ns, lbl=lbl)
+    async def slow_verify(ctx, files, title, gen, q, instructions="", ns="", lbl="", sources=None):
+        await asyncio.sleep(0.8)  # keeps one card in `verify` while the outline fires
+        return await base_verify(ctx, files, title, gen, q, instructions, ns=ns, lbl=lbl, sources=sources)
 
     base_outline = ba._outline_block
 
@@ -732,7 +730,7 @@ async def test_outline_runs_before_artefacts_finish(board_env, monkeypatch):
         snapshot["unfinished"] = sum(1 for c in cards if c["stage"] != "done_artefact")
         return await base_outline(ctx, set_p, files, entries, instructions)
 
-    monkeypatch.setattr(ba, "_levels_block", slow_levels)
+    monkeypatch.setattr(ba, "_verify_block", slow_verify)
     monkeypatch.setattr(ba, "_outline_block", spy_outline)
     ok = await asyncio.wait_for(
         bi.run_boards(ctx, lambda *a, **k: None, files, {"type": "thema"}, None, "", research=False),
@@ -750,7 +748,7 @@ async def test_outline_facts_from_payloads(board_env, tmp_path):
     db, ctx, files = board_env
     await db.kanban_upsert_card(TOPIC, B, "b-1", "block", "done_block",
                                 {"title": "Alpha", "description": "d"})
-    await db.kanban_upsert_card(TOPIC, "artefacts", "alpha", "ablock", "levels",
+    await db.kanban_upsert_card(TOPIC, "artefacts", "alpha", "ablock", "verify",
                                 {"title": "Alpha", "facts": {"Alpha": {"sub eins": {
                                     "sub": "Sub Eins", "prerequisites": "Beta zuerst"}}}})
     await db.kanban_upsert_card(TOPIC, "artefacts", "outline", "outline", "outline",
@@ -764,16 +762,16 @@ async def test_outline_facts_from_payloads(board_env, tmp_path):
 
 async def test_card_view_stepper(testdb):
     """Aktive artefacts-Karte mit Step-Name → step_i/step_n; Alt-String bleibt tolerierbar."""
-    r = {"board": "artefacts", "card_id": "alpha", "stage": "facts", "retries": 0,
+    r = {"board": "artefacts", "card_id": "alpha", "stage": "verify", "retries": 0,
          "payload": {"title": "Alpha"}}
-    v = bi._card_view(r, {"artefacts:alpha"}, {"artefacts:alpha": {"msg": "Facts check 1/2…", "step": "Facts check"}})
-    assert v["status"] == "active" and v["info"] == "Facts check 1/2…"
-    assert v["step_i"] == 2 and v["step_n"] == 3 and v["steps"][0] == "Facts find"
+    v = bi._card_view(r, {"artefacts:alpha"}, {"artefacts:alpha": {"msg": "Fix 1/2…", "step": "Fix"}})
+    assert v["status"] == "active" and v["info"] == "Fix 1/2…"
+    assert v["step_i"] == 2 and v["step_n"] == 2 and v["steps"][0] == "Verify"
     # legacy plain-string live info → no stepper, no crash
-    v2 = bi._card_view(r, {"artefacts:alpha"}, {"artefacts:alpha": "Facts find 0/1…"})
-    assert v2["info"] == "Facts find 0/1…" and "step_n" not in v2
+    v2 = bi._card_view(r, {"artefacts:alpha"}, {"artefacts:alpha": "Verify 0/1…"})
+    assert v2["info"] == "Verify 0/1…" and "step_n" not in v2
     # step outside the card's stage group (e.g. supplement note) → no stepper
-    v3 = bi._card_view(r, {"artefacts:alpha"}, {"artefacts:alpha": {"msg": "x", "step": "Subblocks find"}})
+    v3 = bi._card_view(r, {"artefacts:alpha"}, {"artefacts:alpha": {"msg": "x", "step": "Generate"}})
     assert "step_n" not in v3
 
 
@@ -808,17 +806,17 @@ async def test_seed_map_resolves_cascade(testdb):
 
 
 def test_per_block_functions_accept_wrapper_kwargs():
-    """Die board_artefacts-Wrapper übergeben ns/lbl (subblocks auch seeds) — ein fehlender
-    Parameter stirbt sonst erst im Echt-Lauf als TypeError (Fakes verdecken die Signatur)."""
+    """Die board_artefacts-Prozessoren übergeben ns/lbl (generate auch seeds/sources) —
+    ein fehlender Parameter stirbt sonst erst im Echt-Lauf als TypeError (Fakes verdecken
+    die Signatur)."""
     import inspect
-    import blocks as blx
-    for fn in ("_subblocks_block", "_facts_block", "_levels_block", "_relevance_block",
-               "_question_pattern_block", "_artefacts_block"):
-        params = inspect.signature(getattr(blx, fn)).parameters
+    import block_calls as bc
+    for fn in ("_generate_block", "_verify_block", "_artefakte_block"):
+        params = inspect.signature(getattr(bc, fn)).parameters
         assert "ns" in params and "lbl" in params, fn
-    assert "seeds" in inspect.signature(blx._subblocks_block).parameters
-    for fn in ("_subblocks_block", "_facts_block"):  # Board 2 reicht die Block-Quellen durch
-        assert "sources" in inspect.signature(getattr(blx, fn)).parameters, fn
+    assert "seeds" in inspect.signature(bc._generate_block).parameters
+    for fn in ("_generate_block", "_verify_block"):  # Board 2 reicht die Block-Quellen durch
+        assert "sources" in inspect.signature(getattr(bc, fn)).parameters, fn
 
 
 # ── Inventar-Härtung: Sanitizer, Akronym-Regel, Supplement-Beleg ─────────────────────
@@ -906,7 +904,7 @@ async def _run_flow(ctx, files, timeout=30, **kw):
 
 
 async def test_qa_gate_pauses_on_bad_note(board_env, monkeypatch):
-    """Note unter Schwelle → Flow endet sauber, Board-2-Karten warten in subblocks."""
+    """Note unter Schwelle → Flow endet sauber, Board-2-Karten warten in generate."""
     import qa as qa_mod
     db, ctx, files = board_env
 
@@ -918,7 +916,7 @@ async def test_qa_gate_pauses_on_bad_note(board_env, monkeypatch):
     await _seed(db)
     ok = await _run_flow(ctx, files)
     assert ok
-    warten = await db.kanban_cards(TOPIC, board="artefacts", stage="subblocks")
+    warten = await db.kanban_cards(TOPIC, board="artefacts", stage="generate")
     assert len(warten) == 4  # alle Blöcke gespawnt, keiner verarbeitet
     assert await db.kanban_count(TOPIC, "done_artefact", board="artefacts") == 0
 
@@ -981,7 +979,7 @@ def test_qa_view_pausiert_logic(tmp_path, monkeypatch):
     (tmp_path / TOPIC).mkdir()
     (tmp_path / TOPIC / "r1.json").write_text(_json.dumps(
         {"note": 5.0, "quoten": {"fremd": 0.2}, "fremd": ["X"], "unecht": ["Y"]}), encoding="utf-8")
-    counts = {"inventory": {"done_block": 3}, "artefacts": {"subblocks": 4}}
+    counts = {"inventory": {"done_block": 3}, "artefacts": {"generate": 4}}
     v = bi._qa_view(TOPIC, counts, None)
     assert v["pausiert"] is True and v["note"] == 5.0 and v["befunde"] == ["X", "Y"]
     from types import SimpleNamespace
@@ -1115,11 +1113,11 @@ def test_hat_anker_ziffern_suffix():
     assert not bi._hat_anker("Algorithmus Verfahren", ctoks)  # nur Stopwörter → kein Anker
 
 
-async def test_reset_subblocks_loescht_globale_dateien(testdb, tmp_path):
-    """Reset auf Spalte subblocks: DB-Spiegel UND globale Sidecar-Dateien + ab-*-Resume-Slots
+async def test_reset_generate_loescht_globale_dateien(testdb, tmp_path):
+    """Reset auf Spalte generate: DB-Spiegel UND globale Sidecar-Dateien + ab-*-Resume-Slots
     weg — Reste des Vor-Laufs würden sonst in den frischen Lauf zurückmergen."""
     db = testdb
-    await db.kanban_upsert_card(TOPIC, "artefacts", "alpha", "ablock", "artefacts", {"title": "Alpha"})
+    await db.kanban_upsert_card(TOPIC, "artefacts", "alpha", "ablock", "artefakte", {"title": "Alpha"})
     await db.put_subblock(TOPIC, "alpha", "s1", "Alpha", "S1", status="consensus")
     arbeit = tmp_path / "arbeit"
     (arbeit / "ab-alpha").mkdir(parents=True)
@@ -1128,7 +1126,7 @@ async def test_reset_subblocks_loescht_globale_dateien(testdb, tmp_path):
     for k in ("sidecar", "facts", "sub_roh", "question_pattern", "artefakte"):
         files[k] = tmp_path / f"{k}.json"
         files[k].write_text("{}", encoding="utf-8")
-    moved = await bi.reset_board_from_stage(TOPIC, "artefacts", "subblocks", files)
+    moved = await bi.reset_board_from_stage(TOPIC, "artefacts", "generate", files)
     assert moved == 1
     assert not await db.list_subblocks(TOPIC)
     assert not (arbeit / "ab-alpha").exists()

backend/tests/test_e2e_fake.py

@@ -84,8 +84,9 @@ async def test_e2e_rerun_idempotent(fake_welt, testdb, tmp_path):
 
 @pytest.mark.parametrize("stoerung", [
     {"muster": r"-sub-crossblock-.*-j1$", "modus": "fehler", "mal": 3},   # Ersatzrichter jE
-    {"muster": r"-sub-konsolidierung-.*-j1$", "modus": "garbage", "mal": 1},  # Retry heilt
-    {"muster": r"-facts-c\d+$", "modus": "fehler", "mal": 1},             # Slot-Restart
+    {"muster": r"-sb-verify-.*-j1$", "modus": "garbage", "mal": 1},       # Ersatz-Richter jE
+    {"muster": r"-sb-gen-.*-g1$", "modus": "fehler", "mal": 1},           # degraded: 1 Generator
+    {"muster": r"-art-gen-.*-t1$", "modus": "fehler", "mal": 1},          # Slot-Restart
     {"muster": r"-research-2$", "modus": "fehler", "mal": 3},             # 1 Producer tot
 ])
 async def test_e2e_stoerungen_flow_endet(fake_welt, testdb, tmp_path, stoerung):

backend/tests/test_events.py

@@ -156,7 +156,7 @@ async def test_restart_artefact_card_wipes_only_that_block(testdb):
         await db.upsert_question_pattern(TOPIC, norm, "s1", norm.title(), "Sub Eins", "Frage?")
         await db.put_sub_artifact(TOPIC, norm, "s1", "flashcard", norm.title(), "Sub Eins", "{}")
     assert await bi.restart_artefact_card(TOPIC, "alpha") is True
-    assert (await db.kanban_get_card(TOPIC, "artefacts", "alpha"))["stage"] == "subblocks"
+    assert (await db.kanban_get_card(TOPIC, "artefacts", "alpha"))["stage"] == "generate"
     assert await db.list_subblocks(TOPIC, "alpha") == []
     assert len(await db.list_subblocks(TOPIC, "beta")) == 1  # untouched
     assert await bi.restart_artefact_card(TOPIC, "gibtsnicht") is False

backend/tests/test_konsolidierung.py

@@ -1,10 +1,10 @@
-"""Sub-Konsolidierung: In-Block-Panel (blocks._konsolidiere_subblocks) und
-Cross-Block-Barrier (board_artefacts._proc_konsolidierung) — Judges gefaked, gegen Test-DB."""
+"""Cross-Block-Konsolidierung (board_artefacts._proc_konsolidierung) und Finalize —
+Judges gefaked, gegen Test-DB. Die In-Block-Konsolidierung lebt seit dem Verschmelzungs-
+Umbau in block_calls._verify_block und wird in tests/test_block_calls.py getestet."""
 
 import json
 
 import numpy as np
-import pytest
 
 import blocks
 import board_artefacts as ba
@@ -39,218 +39,6 @@ async def _seed_block(db, bnorm, subs):
         await db.put_subblock(TOPIC, bnorm, blocks._norm_title(s), bnorm.title(), s, status="consensus")
 
 
-# ── In-Block ────────────────────────────────────────────────────────────────────────
-
-async def test_merge_on_unanimity(testdb, tmp_path, monkeypatch):
-    """Beide Judges gruppieren 1+2 → Gewinner (mehr key_points) bleibt, facts-Union,
-    Verlierer wird DB-variant und fliegt aus raw/facts_map."""
-    db = testdb
-    subs = ["Durchstreichung: ~~text~~", "Durchstreichung: ~~text~~ streicht Text durch", "Fett: **text**"]
-    await _seed_block(db, "betonung", subs)
-    raw = {"Betonung": list(subs)}
-    facts = {"Betonung": {
-        blocks._norm_title(subs[0]): {"key_points": ["kp-a"], "cited_facts": [{"text": "z1"}]},
-        blocks._norm_title(subs[1]): {"key_points": ["kp-b", "kp-c"], "cited_facts": [{"text": "z1"}, {"text": "z2"}]},
-    }}
-    fake = _fake_slot({"j1": {"gruppen": [[1, 2]], "luecken": ["Marker-Escaping fehlt"]},
-                       "j2": {"gruppen": [[2, 1]], "luecken": ["Escaping von Markern"]}})
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, facts)
-    assert raw["Betonung"] == [subs[1], "Fett: **text**"]  # Gewinner: 2 key_points > 1
-    wf = facts["Betonung"][blocks._norm_title(subs[1])]
-    assert wf["key_points"] == ["kp-b", "kp-c", "kp-a"]
-    assert wf["cited_facts"] == [{"text": "z1"}, {"text": "z2"}]  # Union ohne Doppel
-    assert blocks._norm_title(subs[0]) not in facts["Betonung"]
-    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "betonung")}
-    assert rows[blocks._norm_title(subs[0])] == "variant"
-    assert rows[blocks._norm_title(subs[1])] == "consensus"
-    journale = list(tmp_path.glob("sub-konsolidierung-*.json"))
-    j = json.loads([p for p in journale if "-j" not in p.stem][0].read_text())
-    # Lücken-Schnitt: Token-Überlappung beider Judges, Formulierung von j1 gewinnt
-    assert j["gruppen"][0]["behalten"] == subs[1] and j["luecken"] == ["Marker-Escaping fehlt"]
-
-
-async def test_dissent_keeps_everything(testdb, tmp_path, monkeypatch):
-    """Nur ein Judge gruppiert → keine Einstimmigkeit → kein Merge."""
-    db = testdb
-    subs = ["Eintrag eins", "Eintrag zwei"]
-    await _seed_block(db, "block", subs)
-    raw = {"Block": list(subs)}
-    fake = _fake_slot({"j1": {"gruppen": [[1, 2]], "luecken": []},
-                       "j2": {"gruppen": [], "luecken": []}})
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, {})
-    assert raw["Block"] == subs
-    assert all(r["status"] == "consensus" for r in await db.list_subblocks(TOPIC, "block"))
-
-
-async def test_judge_failure_fail_open(testdb, tmp_path, monkeypatch):
-    """Ein Judge UND der Ersatz ohne Ergebnis → fail-open, nichts ändert sich."""
-    db = testdb
-    subs = ["Eintrag eins", "Eintrag zwei"]
-    await _seed_block(db, "block", subs)
-    raw = {"Block": list(subs)}
-    fake = _fake_slot({"j1": {"gruppen": [[1, 2]], "luecken": []}})  # j2 UND j3 → FAILED
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, {})
-    assert raw["Block"] == subs
-    assert len(fake.calls) == 3  # j1, j2, Ersatz j3
-
-
-async def test_ersatzrichter_bei_ausfall(testdb, tmp_path, monkeypatch):
-    """j1 fällt aus → Ersatz j3 springt ein; Einstimmigkeit j2+j3 faltet.
-    Vorher entwertete EIN Timeout die gute Stimme (13 Links-Dubletten überlebten)."""
-    db = testdb
-    subs = ["Kurz", "Deutlich längerer Eintrag"]
-    await _seed_block(db, "block", subs)
-    raw = {"Block": list(subs)}
-    fake = _fake_slot({"j2": {"gruppen": [[1, 2]], "luecken": []},
-                       "j3": {"gruppen": [[2, 1]], "luecken": []}})  # j1 → FAILED
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, {})
-    assert raw["Block"] == ["Deutlich längerer Eintrag"]
-
-
-async def test_negation_guard_blocks_merge(testdb, tmp_path, monkeypatch):
-    """Gegensätzliche Aussagen werden selbst bei einstimmigen Judges nicht gefaltet."""
-    db = testdb
-    subs = ["Tabs werden expandiert", "Tabs werden nicht expandiert"]
-    await _seed_block(db, "tabs", subs)
-    raw = {"Tabs": list(subs)}
-    fake = _fake_slot({"j1": {"gruppen": [[1, 2]], "luecken": []},
-                       "j2": {"gruppen": [[1, 2]], "luecken": []}})
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, {})
-    assert raw["Tabs"] == subs
-
-
-async def test_resume_skips_judges(testdb, tmp_path, monkeypatch):
-    """Vorhandene j-Dateien → kein neuer Agenten-Call, Ergebnis wird übernommen."""
-    db = testdb
-    subs = ["Eintrag eins", "Eintrag zwei lang"]
-    await _seed_block(db, "block", subs)
-    raw = {"Block": list(subs)}
-    import hashlib
-    h = hashlib.md5("\n".join(subs).encode()).hexdigest()[:8]
-    for j in (1, 2):
-        (tmp_path / f"sub-konsolidierung-{h}-j{j}.json").write_text(
-            json.dumps({"gruppen": [[1, 2]], "luecken": []}), encoding="utf-8")
-
-    async def kein_agent(*a, **kw):
-        raise AssertionError("Resume darf keinen Agenten starten")
-
-    monkeypatch.setattr(blocks, "run_single_slot", kein_agent)
-    await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, {})
-    assert raw["Block"] == ["Eintrag zwei lang"]
-
-
-def test_schema_accepts_both_group_forms():
-    """Alte Listenform [1,4] und neue {haupt, weitere}-Form parsen beide; kataloge/fremd optional."""
-    alt = blocks._konsolidierung_schema({"gruppen": [[1, 4]], "luecken": []}, 5)
-    assert alt["gruppen"] == [{"haupt": None, "ids": [1, 4]}] and alt["fremd"] == set()
-    neu = blocks._konsolidierung_schema(
-        {"gruppen": [{"haupt": 4, "weitere": [1]}],
-         "kataloge": [{"titel": "Katalog: Symbole", "mitglieder": [2, 3]}],
-         "fremd": [5], "luecken": ["x"]}, 5)
-    assert neu["gruppen"] == [{"haupt": 4, "ids": [1, 4]}]
-    assert neu["kataloge"] == [{"titel": "Katalog: Symbole", "ids": [2, 3]}]
-    assert neu["fremd"] == {5} and neu["luecken"] == ["x"]
-    assert blocks._konsolidierung_schema({"gruppen": [{"haupt": 9, "weitere": [1]}]}, 5) == \
-        {"gruppen": [], "kataloge": [], "fremd": set(), "luecken": []}  # id out of range
-
-
-async def test_haupt_beats_heuristic(testdb, tmp_path, monkeypatch):
-    """Judges nennen den kürzeren Eintrag als haupt → er gewinnt trotz weniger key_points."""
-    db = testdb
-    subs = ["Basis", "Detailregel mit sehr langem Titel und Facts"]
-    await _seed_block(db, "block", subs)
-    raw = {"Block": list(subs)}
-    facts = {"Block": {blocks._norm_title(subs[1]): {"key_points": ["a", "b", "c"]}}}
-    fake = _fake_slot({"j1": {"gruppen": [{"haupt": 1, "weitere": [2]}], "luecken": []},
-                       "j2": {"gruppen": [{"haupt": 1, "weitere": [2]}], "luecken": []}})
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, facts)
-    assert raw["Block"] == ["Basis"]
-    assert facts["Block"][blocks._norm_title("Basis")]["key_points"] == ["a", "b", "c"]  # Union geerbt
-
-
-async def test_katalog_bundles_to_new_row(testdb, tmp_path, monkeypatch):
-    """Einstimmige Katalog-Mitglieder → neue consensus-Zeile mit Facts-Union, Mitglieder variant."""
-    db = testdb
-    subs = ["Pfeilsymbole: a b c", "Mengensymbole: d e f", "Eigene Regel"]
-    await _seed_block(db, "mathe", subs)
-    raw = {"Mathe": list(subs)}
-    facts = {"Mathe": {blocks._norm_title(subs[0]): {"key_points": ["kp1"]},
-                       blocks._norm_title(subs[1]): {"key_points": ["kp2"]}}}
-    kat = {"titel": "Symbolkatalog: Pfeile und Mengen", "mitglieder": [1, 2]}
-    fake = _fake_slot({"j1": {"gruppen": [], "kataloge": [kat], "luecken": []},
-                       "j2": {"gruppen": [], "kataloge": [kat], "luecken": []}})
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, facts)
-    assert raw["Mathe"] == ["Eigene Regel", "Symbolkatalog: Pfeile und Mengen"]
-    kn = blocks._norm_title("Symbolkatalog: Pfeile und Mengen")
-    assert sorted(facts["Mathe"][kn]["key_points"]) == ["kp1", "kp2"]
-    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "mathe")}
-    assert rows[kn] == "consensus"
-    assert rows[blocks._norm_title(subs[0])] == "variant"
-
-
-async def test_katalog_dissent_keeps_members(testdb, tmp_path, monkeypatch):
-    db = testdb
-    subs = ["Pfeilsymbole: a b c", "Mengensymbole: d e f"]
-    await _seed_block(db, "mathe", subs)
-    raw = {"Mathe": list(subs)}
-    fake = _fake_slot({"j1": {"gruppen": [], "kataloge": [{"titel": "K", "mitglieder": [1, 2]}], "luecken": []},
-                       "j2": {"gruppen": [], "kataloge": [], "luecken": []}})
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, {})
-    assert raw["Mathe"] == subs
-
-
-async def test_fremd_unanimous_discards(testdb, tmp_path, monkeypatch):
-    """Einstimmig fremd → discarded + raus; einseitig fremd → bleibt."""
-    db = testdb
-    subs = ["CSS display überschreibt Verhalten", "Echte Markdown-Regel", "Nur einer hält es für fremd"]
-    await _seed_block(db, "block", subs)
-    raw = {"Block": list(subs)}
-    fake = _fake_slot({"j1": {"gruppen": [], "fremd": [1, 3], "luecken": []},
-                       "j2": {"gruppen": [], "fremd": [1], "luecken": []}})
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    luecken = await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, {})
-    assert raw["Block"] == [subs[1], subs[2]]
-    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "block")}
-    assert rows[blocks._norm_title(subs[0])] == "discarded"
-    assert rows[blocks._norm_title(subs[2])] == "consensus"
-    assert luecken == {}
-
-
-async def test_luecken_nur_bei_einstimmigkeit(testdb, tmp_path, monkeypatch):
-    """Nur Lücken mit Token-Überlappung BEIDER Judges überleben; einseitige fallen weg."""
-    db = testdb
-    subs = ["Eintrag eins", "Eintrag zwei"]
-    await _seed_block(db, "block", subs)
-    raw = {"Block": list(subs)}
-    fake = _fake_slot({"j1": {"gruppen": [], "luecken": ["Inline-HTML fehlt", "Front-Matter"]},
-                       "j2": {"gruppen": [], "luecken": ["nichts zu Inline-HTML"]}})
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    luecken = await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, {})
-    assert luecken == {"Block": ["Inline-HTML fehlt"]}
-
-
-async def test_kp_deckel_im_judge_prompt(testdb, tmp_path, monkeypatch):
-    """Prompt zeigt max. 3 key_points je Sub (Timeout-Schutz); die Union bleibt voll."""
-    db = testdb
-    subs = ["Eintrag eins", "Eintrag zwei"]
-    await _seed_block(db, "block", subs)
-    raw = {"Block": list(subs)}
-    facts = {"Block": {blocks._norm_title(subs[0]): {"key_points": [f"kp{i}" for i in range(1, 6)]}}}
-    fake = _fake_slot({"j1": {"gruppen": [], "luecken": []}, "j2": {"gruppen": [], "luecken": []}})
-    monkeypatch.setattr(blocks, "run_single_slot", fake)
-    await blocks._konsolidiere_subblocks(_ctx(), {"arbeit": tmp_path}, raw, facts)
-    prompt = fake.calls[0]["prompt"]
-    assert "kp3" in prompt and "kp4" not in prompt
-
-
 def test_luecken_schnitt_cap():
     l1 = [f"Aspekt-{k} fehlt" for k in ("eins", "zwei", "drei", "vier", "fünf")]
     assert blocks._luecken_schnitt(l1, list(l1)) == l1[:3]  # Cap 3
@@ -266,91 +54,6 @@ def test_neg_set_lemmatisiert():
     assert blocks._neg_set("niemals gerendert") == blocks._neg_set("nie gerendert")
 
 
-# ── Lücken-Nachfass ─────────────────────────────────────────────────────────────────
-
-async def _nachfass_env(db, monkeypatch, facts_result):
-    subs = ["Eintrag eins"]
-    await _seed_block(db, "block", subs)
-    raw = {"Block": list(subs)}
-    facts_map = {"Block": {}}
-
-    async def fake_race(topic, label, slots, quorum, timeout, provider, cancelled=None, grace=0):
-        return [{"Block": ["Eintrag eins", "Neuer Aspekt"]}]
-
-    async def fake_facts(ctx, set_p, files, fraw, q, folder, instructions, ns="", lbl="", sources=None, slim=False):
-        assert slim is True  # Nachfass nutzt die schlanke Facts-Variante
-        assert list(fraw["Block"]) == ["Neuer Aspekt"]  # nur der frische Fund geht ins Gate
-        return facts_result
-
-    monkeypatch.setattr(blocks, "_race", fake_race)
-    monkeypatch.setattr(blocks, "_facts_block", fake_facts)
-    monkeypatch.setattr(blocks, "EMBEDDING_AKTIV", False)
-    return raw, facts_map
-
-
-async def test_nachfass_adopts_backed_find(testdb, tmp_path, monkeypatch):
-    db = testdb
-    nn = blocks._norm_title("Neuer Aspekt")
-    raw, facts_map = await _nachfass_env(db, monkeypatch,
-                                         ({"Block": {nn: {"key_points": ["kp"]}}}, {}))
-    n = await blocks._luecken_runde(_ctx(), {"arbeit": tmp_path}, "Block", ["Aspekt"],
-                                    raw, facts_map, {"type": "thema"}, None)
-    assert n == 1 and raw["Block"] == ["Eintrag eins", "Neuer Aspekt"]
-    assert facts_map["Block"][nn]["key_points"] == ["kp"]
-    rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "block")}
-    assert rows[nn] == "consensus"
-
-
-async def test_nachfass_drops_unbacked_find(testdb, tmp_path, monkeypatch):
-    db = testdb
-    nn = blocks._norm_title("Neuer Aspekt")
-    raw, facts_map = await _nachfass_env(db, monkeypatch, ({}, {"Block": {nn}}))
-    n = await blocks._luecken_runde(_ctx(), {"arbeit": tmp_path}, "Block", ["Aspekt"],
-                                    raw, facts_map, {"type": "thema"}, None)
-    assert n == 0 and raw["Block"] == ["Eintrag eins"]
-    assert not any(r["sub_norm"] == nn for r in await db.list_subblocks(TOPIC, "block"))
-
-
-async def test_nachfass_drops_find_without_facts(testdb, tmp_path, monkeypatch):
-    """HARTES Gate: kein Facts-Eintrag = kein Beleg = keine Übernahme — nicht nur
-    aktiv Verworfenes fliegt (Bilder-Lauf: 13 von 18 kamen ohne Beleg durch)."""
-    db = testdb
-    raw, facts_map = await _nachfass_env(db, monkeypatch, ({}, {}))  # Facts fand NICHTS
-    n = await blocks._luecken_runde(_ctx(), {"arbeit": tmp_path}, "Block", ["Aspekt"],
-                                    raw, facts_map, {"type": "thema"}, None)
-    assert n == 0 and raw["Block"] == ["Eintrag eins"]
-
-
-async def test_facts_stage_konsolidiert_nachfass_funde_erneut(testdb, tmp_path, monkeypatch):
-    """Kreis geschlossen: nach Übernahmen läuft die Konsolidierung ein zweites Mal;
-    deren Lücken lösen KEINEN weiteren Nachfass aus."""
-    db = testdb
-    payload = {"title": "Alpha", "raw": {"Alpha": ["s1"]}}
-    await db.kanban_upsert_card(TOPIC, "artefacts", "alpha", "ablock", "facts", payload)
-    calls = {"kons": 0, "nf": 0}
-
-    async def fake_facts(ctx, set_p, files, raw, q, folder, instructions, ns="", lbl="", sources=None, slim=False):
-        return {"Alpha": {}}, {}
-
-    async def fake_kons(ctx, files, raw, facts_map, instructions="", ns="", lbl=""):
-        calls["kons"] += 1
-        return {"Alpha": ["Lücke X"]}  # meldet auch in Runde 2 — darf nicht erneut nachfassen
-
-    async def fake_nf(ctx, files, title, luecken, raw, facts_map, q, folder,
-                      instructions="", ns="", lbl="", sources=None):
-        calls["nf"] += 1
-        return 2
-
-    monkeypatch.setattr(ba, "_facts_block", fake_facts)
-    monkeypatch.setattr(ba, "_konsolidiere_subblocks", fake_kons)
-    monkeypatch.setattr(ba, "_luecken_runde", fake_nf)
-    flow = Flow(TOPIC, work_dir=tmp_path)
-    await ba._proc_facts(_ctx(), flow, {"arbeit": tmp_path}, {"type": "thema"}, None, "",
-                         [{"card_id": "alpha", "payload": payload}])
-    assert calls == {"kons": 2, "nf": 1}
-    assert (await db.kanban_get_card(TOPIC, "artefacts", "alpha"))["stage"] == "levels"
-
-
 async def test_finalize_purges_stale_rows(testdb, tmp_path):
     """Re-Run-Waisen: Finalize löscht Alt-Fragen/-Artefakte des Blocks vor dem Upsert."""
     db = testdb
@@ -486,11 +189,13 @@ async def test_crossblock_without_embedding_advances(testdb, tmp_path, monkeypat
         assert (await db.kanban_get_card(TOPIC, "artefacts", cid))["stage"] == ba.DONE
 
 
-async def test_crossblock_nachzuegler_zurueck_zu_fragen(testdb, tmp_path, monkeypatch):
-    """Resume-Karte aus der alten Stage-Position (kein pattern im Payload) → zurück nach
-    question_pattern, KEIN Dedup — finalize würde den Fold sonst re-spiegeln."""
+async def test_crossblock_nachzuegler_zurueck_zum_erzeugen(testdb, tmp_path, monkeypatch):
+    """Resume-Karte ohne pattern im Payload → zurück nach generate (bzw. artefakte bei
+    vorhandenem sidecar), KEIN Dedup — finalize würde den Fold sonst re-spiegeln."""
     db = testdb
     flow, cards, files = await _cross_env(db, tmp_path, finalisiert=False)
+    cards[1]["payload"]["sidecar"] = {"Beta": []}  # hat Verify schon hinter sich
+    await db.kanban_set_payload(TOPIC, "artefacts", "beta", cards[1]["payload"])
 
     async def kein_agent(*a, **kw):
         raise AssertionError("Nachzügler dürfen keinen Dedup auslösen")
@@ -498,8 +203,8 @@ async def test_crossblock_nachzuegler_zurueck_zu_fragen(testdb, tmp_path, monkey
     monkeypatch.setattr(ba, "embedding", _FakeEmb)
     monkeypatch.setattr(ba, "run_single_slot", kein_agent)
     await ba._proc_konsolidierung(_ctx(), flow, files, "", cards)
-    for cid in ("alpha", "beta"):
-        assert (await db.kanban_get_card(TOPIC, "artefacts", cid))["stage"] == "question_pattern"
+    assert (await db.kanban_get_card(TOPIC, "artefacts", "alpha"))["stage"] == "generate"
+    assert (await db.kanban_get_card(TOPIC, "artefacts", "beta"))["stage"] == "artefakte"
     rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "beta")}
     assert rows[blocks._norm_title("Gleiche Aussage")] == "consensus"
 
@@ -558,47 +263,3 @@ async def test_crossblock_chunking_faltet_global(testdb, tmp_path, monkeypatch):
     assert len(fake.calls) == 4  # 2 Chunks × j1/j2
     rows = {r["sub_norm"]: r["status"] for r in await db.list_subblocks(TOPIC, "beta")}
     assert set(rows.values()) == {"variant"}  # beide Dubletten global gefaltet
-
-
-async def test_facts_nachfass_holt_nur_fehlende(testdb, tmp_path, monkeypatch):
-    """Nachfass ruft den slim-Facts-Lauf NUR mit den facts-losen Subs und merged die Funde;
-    Vorhandenes bleibt unberührt, Subs werden nie verworfen."""
-    gesehen = {}
-
-    async def fake_facts_block(ctx, set_p, files, raw, q, folder, instructions, ns="", lbl="",
-                               sources=None, slim=False):
-        gesehen["raw"] = raw
-        gesehen["slim"] = slim
-        return ({"Alpha": {"ohne beleg": {"key_points": ["kp neu"]},
-                           "mit beleg": {"key_points": ["DARF NICHT GEWINNEN"]}}},
-                {"Alpha": {"ohne beleg"}})  # discard-Urteil wird ignoriert
-
-    monkeypatch.setattr(blocks, "_facts_block", fake_facts_block)
-    raw = {"Alpha": ["Mit Beleg", "Ohne Beleg"]}
-    facts_map = {"Alpha": {"mit beleg": {"key_points": ["kp alt"]}}}
-    ctx = GenContext(topic=TOPIC, provider="claude", is_cancelled=lambda: False)
-    n = await blocks._facts_nachfass(ctx, {"arbeit": tmp_path}, raw, facts_map, {}, None)
-    assert n == 1
-    assert gesehen["slim"] and gesehen["raw"] == {"Alpha": ["Ohne Beleg"]}
-    assert facts_map["Alpha"]["ohne beleg"]["key_points"] == ["kp neu"]
-    assert facts_map["Alpha"]["mit beleg"]["key_points"] == ["kp alt"]
-    assert raw["Alpha"] == ["Mit Beleg", "Ohne Beleg"]  # kein Verwurf
-
-
-async def test_levels_merge_fuzzy_match(testdb, tmp_path, monkeypatch):
-    """Levels-Agent paraphrasiert den Sub-Titel → facts hängen trotzdem am Sidecar-Eintrag
-    (eindeutiger Präfix-Match statt stillem Grounding-Verlust)."""
-    db = testdb
-    payload = {"title": "Alpha", "raw": {"Alpha": ["Marker Regel: Details dazu"]},
-               "facts": {"Alpha": {blocks._norm_title("Marker Regel: Details dazu"):
-                                   {"key_points": ["kp"]}}}}
-    await db.kanban_upsert_card(TOPIC, "artefacts", "alpha", "ablock", "levels", payload)
-    cards = [{"card_id": "alpha", "payload": payload}]
-
-    async def fake_levels_block(ctx, set_p, files, raw, instructions, ns="", lbl=""):
-        return {"Alpha": [{"title": "Marker Regel", "level": "beginner"}]}  # gekürzter Titel
-
-    monkeypatch.setattr(ba, "_levels_block", fake_levels_block)
-    await ba._proc_levels(_ctx(), Flow(TOPIC, work_dir=tmp_path), {"arbeit": tmp_path}, "", cards)
-    card = await db.kanban_get_card(TOPIC, "artefacts", "alpha")
-    assert card["payload"]["sidecar"]["Alpha"][0]["facts"] == {"key_points": ["kp"]}

backend/tests/test_subblocks.py

@@ -1,16 +1,14 @@
-"""Subbaustein-Qualität: Varianten-Konsens, Seed-Garantie, Nachfass, Outline-Review."""
+"""Subbaustein-Helfer: Varianten-Cluster, Evidence-Packs, Outline-Review, Hash/Key-Auflösung.
+Die verschmolzene Block-Pipeline (Generate/Verify/Artefakte) testet tests/test_block_calls.py."""
 
 import json
-import re
 
 import numpy as np
-import pytest
 
 import blocks as blx
 from pipeline import GenContext
 
 TOPIC = "t"
-_MD_PATH = re.compile(r"(/\S+\.md)")
 
 
 # ── _variant_clusters (pure) ─────────────────────────────────────────────────────────
@@ -37,157 +35,6 @@ def test_variant_clusters_negation_guard():
     assert len(cl) == 2  # antonyms never merge, no matter the cosine
 
 
-# ── _subblocks_block integration (fake race + fake embeddings) ──────────────────────
-
-def _fake_sims(texts):
-    """Markertoken matrix: same first word → 0.95, else 0."""
-    n = len(texts)
-    m = np.eye(n)
-    key = lambda t: t.split()[0].casefold()
-    for i in range(n):
-        for j in range(n):
-            if i != j and key(texts[i]) == key(texts[j]):
-                m[i][j] = 0.95
-    return m
-
-
-def _mk_race(finder_by_agent):
-    """Key-routed _race fake. Finder round 1 → scripted per-agent subs; later finder and
-    catch-up rounds → nothing; clarify judges echo the consensus lines from their prompt."""
-    prompts = []
-
-    async def fake_race(topic, label, slots, quorum, timeout, provider, on_update=None,
-                        cancelled=None, *, grace=None, min_runtime=None, max_runtime=None, late=None):
-        outs = []
-        for slot in slots:
-            key, prompt = slot["key"], slot["prompt"]
-            prompts.append((key, prompt))
-            fake_race.slots_seen.append(slot)
-            if "-subblock-final-" in key:
-                # no-tool judges reply as TEXT; the payload sink writes the j-file itself
-                kons = re.search(r"Konsens \(≥2 finders\):\n(.*?)\nUnsicher", prompt, re.S)
-                subs = [l[2:] for l in (kons.group(1).splitlines() if kons else [])
-                        if l.startswith("- ") and l != "- (keiner)"]
-                if subs:
-                    text = "<!-- block: Alpha -->\n" + "\n".join(f"- {s}" for s in subs)
-                    outs.append(slot["payload"]((0, text, "")))
-                continue
-            if "-r1-" in key:
-                agent = int(key.rsplit("-", 1)[1])
-                subs = finder_by_agent.get(agent) or []
-                if subs:  # Finder antworten als TEXT (Marker-Format), kein out_path mehr
-                    text = "<!-- block: Alpha -->\n" + "\n".join(f"- {s}" for s in subs)
-                    outs.append(slot["payload"]((0, text, "")))
-        outs = [o for o in outs if o]
-        return outs or None
-    fake_race.slots_seen = []
-    return fake_race, prompts
-
-
-@pytest.fixture
-def sub_env(testdb, tmp_path, monkeypatch):
-    monkeypatch.setattr(blx, "EMBEDDING_AKTIV", True)
-    monkeypatch.setattr(blx.embedding, "available", lambda: True)
-    monkeypatch.setattr(blx.embedding, "embed_sims", _fake_sims)
-    ctx = GenContext(topic=TOPIC, provider="claude", is_cancelled=lambda: False)
-    files = {"arbeit": tmp_path}
-    return testdb, ctx, files
-
-
-async def _run(ctx, files, monkeypatch, finder_by_agent, seeds=None):
-    fake, prompts = _mk_race(finder_by_agent)
-    monkeypatch.setattr(blx, "_race", fake)
-    raw = await blx._subblocks_block(ctx, lambda *a, **k: None, files, {1: "Alpha — Grundkonzept"},
-                                     "", wipe=False, ns="x-", seeds=seeds)
-    return raw, prompts
-
-
-async def test_variant_consensus_end_to_end(sub_env, monkeypatch):
-    """3 Einzelfunde in 3 Formulierungen → EIN consensus-Repräsentant; Varianten gehen
-    nicht als „Unsicher" ins Panel."""
-    db, ctx, files = sub_env
-    raw, prompts = await _run(ctx, files, monkeypatch, {
-        1: ["Umbruch braucht Marker"],
-        2: ["Umbruch erfordert explizite Marker!"],
-        3: ["Umbruch verlangt zwei Leerzeichen als Marker"],
-    })
-    assert raw == {"Alpha": ["Umbruch verlangt zwei Leerzeichen als Marker"]}  # longest = rep
-    rows = await db.list_subblocks(TOPIC, "alpha")
-    status = sorted(r["status"] for r in rows)
-    assert status == ["consensus", "variant", "variant"]
-    clarify_prompts = [p for k, p in prompts if "-subblock-final-" in k]
-    assert clarify_prompts and "Umbruch braucht Marker" not in clarify_prompts[0]
-
-
-async def test_seed_promotes_single_find(sub_env, monkeypatch):
-    """Seed deckt einen verworfenen Einzelfund lexikalisch → Promotion zu consensus."""
-    db, ctx, files = sub_env
-    raw, _ = await _run(ctx, files, monkeypatch, {
-        1: ["Alpha Grundlagen", "Zeilenumbruch Regeln im Detail"],
-        2: ["Alpha Grundlagen"],
-    }, seeds=["Zeilenumbruch Regeln"])
-    assert "Zeilenumbruch Regeln im Detail" in raw["Alpha"]
-    row = next(r for r in await db.list_subblocks(TOPIC, "alpha")
-               if r["sub_norm"] == "zeilenumbruch regeln im detail")
-    assert row["status"] == "consensus"
-
-
-async def test_seed_inserted_when_nothing_found(sub_env, monkeypatch):
-    """Seed ohne jeden Fund wird als eigener consensus-Sub eingefügt (Facts-Gate prüft später)."""
-    db, ctx, files = sub_env
-    raw, _ = await _run(ctx, files, monkeypatch, {
-        1: ["Alpha Grundlagen"], 2: ["Alpha Grundlagen"],
-    }, seeds=["Fußnoten Syntax"])
-    assert "Fußnoten Syntax" in raw["Alpha"]
-    row = next(r for r in await db.list_subblocks(TOPIC, "alpha")
-               if r["sub_title"] == "Fußnoten Syntax")
-    assert row["status"] == "consensus"
-
-
-async def test_seed_covered_no_duplicate(sub_env, monkeypatch):
-    """Seed lexikalisch von einem consensus-Sub abgedeckt → nichts eingefügt."""
-    db, ctx, files = sub_env
-    raw, _ = await _run(ctx, files, monkeypatch, {
-        1: ["Tabs werden zu Leerzeichen expandiert"], 2: ["Tabs werden zu Leerzeichen expandiert"],
-    }, seeds=["Tabs"])
-    assert raw == {"Alpha": ["Tabs werden zu Leerzeichen expandiert"]}
-
-
-async def test_wipe_false_is_idempotent(sub_env, monkeypatch):
-    """Zweiter Karten-Lauf kumuliert keine Mentions (per-Block-Wipe)."""
-    db, ctx, files = sub_env
-    await _run(ctx, files, monkeypatch, {1: ["Alpha Grundlagen"], 2: ["Alpha Grundlagen"]})
-    first = {r["sub_norm"]: r["mentions"] for r in await db.list_subblocks(TOPIC, "alpha")}
-    await _run(ctx, files, monkeypatch, {1: ["Alpha Grundlagen"], 2: ["Alpha Grundlagen"]})
-    second = {r["sub_norm"]: r["mentions"] for r in await db.list_subblocks(TOPIC, "alpha")}
-    assert first == second
-
-
-async def test_catchup_adds_and_stops(sub_env, monkeypatch, tmp_path):
-    """Block unter SUBBLOCK_MIN: Nachfass-Runde findet Neues → eigenes Final-File,
-    Konsens wächst; zweite Runde ohne Neues → Ende."""
-    db, ctx, files = sub_env
-    fake, prompts = _mk_race({1: ["Alpha Grundlagen"], 2: ["Alpha Grundlagen"]})
-    base = fake
-    hit = {"n": 0}
-
-    async def with_catchup(topic, label, slots, *a, **k):
-        if any("-subblock-x" in s["key"] for s in slots):
-            hit["n"] += 1
-            if hit["n"] == 1:  # first catch-up round: both agents agree on one new sub
-                text = "<!-- block: Alpha -->\n- Vertiefung der Konzepte"
-                return [slot["payload"]((0, text, "")) for slot in slots[:2]]
-            return None
-        return await base(topic, label, slots, *a, **k)
-
-    monkeypatch.setattr(blx, "_race", with_catchup)
-    raw = await blx._subblocks_block(ctx, lambda *a, **k: None, files, {1: "Alpha — Grundkonzept"},
-                                     "", wipe=False, ns="x-")
-    assert set(raw["Alpha"]) == {"Alpha Grundlagen", "Vertiefung der Konzepte"}
-    assert (tmp_path / "subblock-final-c1-x1.md").exists()
-    assert hit["n"] == 2  # round 2 ran, found nothing, loop ended
-
-
 # ── Outline-Review ───────────────────────────────────────────────────────────────────
 
 def test_outline_review_schema():
@@ -238,52 +85,8 @@ async def test_outline_review_moves_block(testdb, tmp_path, monkeypatch):
     plan2 = await blx._outline_block(ctx, lambda *a, **k: None, files, entries, "")
     assert plan2["chapters"][0]["numbers"] == [1, 2, 6]
 
-async def test_paraphrase_saturation_stops_early(sub_env, monkeypatch):
-    """Runde 2 liefert nur eine Paraphrase → zählt nicht als neu, Schleife endet ohne r3.
-    Die Paraphrase liegt trotzdem in der DB (Mention fürs Cluster-Voting)."""
-    db, ctx, files = sub_env
-    base_fake, prompts = _mk_race({1: ["Umbruch braucht Marker"], 2: ["Umbruch braucht Marker"]})
 
-    async def with_r2(topic, label, slots, *a, **k):
-        if any("-r2-" in s["key"] for s in slots):
-            text = "<!-- block: Alpha -->\n- Umbruch erfordert explizite Marker!"
-            return [slot["payload"]((0, text, "")) for slot in slots[:2]]
-        return await base_fake(topic, label, slots, *a, **k)
-
-    monkeypatch.setattr(blx, "_race", with_r2)
-    raw = await blx._subblocks_block(ctx, lambda *a, **k: None, files, {1: "Alpha — Grundkonzept"},
-                                     "", wipe=False, ns="x-")
-    assert raw["Alpha"]  # Konsens steht
-    assert not any("-r3-" in k for k, _ in prompts)  # Paraphrase hielt die Schleife NICHT am Leben
-    rows = await db.list_subblocks(TOPIC, "alpha")
-    assert any(r["sub_title"] == "Umbruch erfordert explizite Marker!" for r in rows)
-
-
-async def test_round_cap_stops_endless_finders(sub_env, monkeypatch):
-    """Jede Runde ein echt neues Konzept → hartes Cap stoppt bei SUBBLOCK_MAX_ROUNDS."""
-    db, ctx, files = sub_env
-    _, prompts = _mk_race({})
-
-    async def endless(topic, label, slots, *a, **k):
-        if "-subblock-final-" in slots[0]["key"]:
-            return None  # panel fails → consensus fallback
-        outs = []
-        import re as _re
-        rn = _re.search(r"-r(\d+)-", slots[0]["key"])
-        n = rn.group(1) if rn else "x"
-        for slot in slots[:2]:
-            prompts.append((slot["key"], slot["prompt"]))
-            outs.append(slot["payload"]((0, f"<!-- block: Alpha -->\n- Konzept{n} ist eigenständig", "")))
-        return outs
-
-    monkeypatch.setattr(blx, "_race", endless)
-    await blx._subblocks_block(ctx, lambda *a, **k: None, files, {1: "Alpha — Grundkonzept"},
-                               "", wipe=False, ns="x-")
-    max_round = max(int(k.split("-r")[1].split("-")[0]) for k, _ in prompts if "-r" in k and "-subblock-c" in k)
-    assert max_round == blx.SUBBLOCK_MAX_ROUNDS
-
-
-# ── Inline-Evidenz für Judges (Token-Umbau) ──────────────────────────────────────────
+# ── Inline-Evidenz (Evidence-Packs für die verschmolzenen Calls) ─────────────────────
 
 def _corpus(tmp_path):
     d = tmp_path / "korpus"
@@ -348,51 +151,6 @@ def test_sink_json_writes_only_valid(tmp_path):
     assert bad is None and not (tmp_path / "x.json").exists()
 
 
-async def test_clarify_inline_evidence_no_tools(sub_env, monkeypatch, tmp_path):
-    """Mit Korpus: Judges UND Finder bekommen Auszüge inline und laufen ohne Tools
-    (Text-Antwort); die Dateien schreibt die Engine. Der Finder verlor vorher 3–13
-    Tool-Runden pro Call mit der Material-Suche via glob/grep/bash."""
-    db, ctx, files = sub_env
-    d = _corpus(tmp_path)
-    monkeypatch.setattr(blx, "source_folder", lambda t: d)
-    monkeypatch.setattr(blx, "load_source", lambda t: {"type": "uni"})
-    fake, prompts = _mk_race({1: ["Alpha Grundlagen"], 2: ["Alpha Grundlagen"]})
-    monkeypatch.setattr(blx, "_race", fake)
-    raw = await blx._subblocks_block(ctx, lambda *a, **k: None, files, {1: "Alpha — Grundkonzept"},
-                                     "", wipe=False, ns="x-", sources=["Skript.txt"])
-    assert raw == {"Alpha": ["Alpha Grundlagen"]}
-    judges = [s for s in fake.slots_seen if "-subblock-final-" in s["key"]]
-    finders = [s for s in fake.slots_seen if "-r1-" in s["key"]]
-    assert judges and all(s["capabilities"] == "none" for s in judges)
-    assert "── Skript.txt" in judges[0]["prompt"]
-    assert "ls/find" not in judges[0]["prompt"]  # keine Selbst-Recherche-Anweisung mehr
-    assert finders and all(s["capabilities"] == "none" for s in finders)
-    assert "── Skript.txt" in finders[0]["prompt"]  # Auszüge inline statt Dateisystem-Suche
-    assert "web search" not in finders[0]["prompt"]
-    assert list(tmp_path.glob("subblock-final-*-j*.md"))  # Engine persistiert die Judge-Antwort
-
-
-async def test_thema_nutzt_research_material_inline(sub_env, monkeypatch, tmp_path):
-    """thema mit Research-Fundstellen (arbeit/material/*.txt): Finder bekommt sie inline
-    und läuft ohne Tools — vorher eigene Websuche pro Call (Reasoning-Schleifen, Retries)."""
-    db, ctx, files = sub_env
-    monkeypatch.setattr(blx, "source_folder", lambda t: None)
-    md = tmp_path / "arbeit" / "material"
-    md.mkdir(parents=True)
-    (md / "research-1.txt").write_text(
-        "https://example.org/alpha\nAlpha Grundlagen: der Kernbegriff, gut belegt.\n",
-        encoding="utf-8")
-    monkeypatch.setattr(blx, "arbeit_dir", lambda t: tmp_path / "arbeit")
-    fake, prompts = _mk_race({1: ["Alpha Grundlagen"], 2: ["Alpha Grundlagen"]})
-    monkeypatch.setattr(blx, "_race", fake)
-    raw = await blx._subblocks_block(ctx, lambda *a, **k: None, files, {1: "Alpha — Grundkonzept"},
-                                     "", wipe=False, ns="x-")
-    assert raw == {"Alpha": ["Alpha Grundlagen"]}
-    finders = [s for s in fake.slots_seen if "-r1-" in s["key"]]
-    assert finders and all(s["capabilities"] == "none" for s in finders)
-    assert "── research-1.txt" in finders[0]["prompt"]  # Fundstellen inline
-
-
 def test_material_folder_fallbacks(monkeypatch, tmp_path):
     """Echte Quelle gewinnt; sonst arbeit/material mit Inhalt; sonst None."""
     monkeypatch.setattr(blx, "source_folder", lambda t: tmp_path / "quelle")
@@ -407,115 +165,6 @@ def test_material_folder_fallbacks(monkeypatch, tmp_path):
     assert blx.material_folder("t") == md
 
 
-async def test_panel_2of3_kehrt_bei_einigkeit_zurueck(tmp_path):
-    """Zwei übereinstimmende Verdicts → Rückkehr ohne den langsamen Dritten; sein
-    Ergebnis wird detached nachpersistiert (Resume)."""
-    import asyncio as aio
-    gesunken = {}
-
-    async def judge(j, delay, antwort):
-        await aio.sleep(delay)
-        return (0, antwort, "")
-
-    tasks = {aio.create_task(judge(1, 0.01, "a")): 1,
-             aio.create_task(judge(2, 0.02, "a")): 2,
-             aio.create_task(judge(3, 5.0, "b")): 3}
-
-    def sink(j, r):
-        gesunken[j] = r[1]
-
-    import time
-    t0 = time.monotonic()
-    await blx._panel_2of3(tasks, sink, lambda: list(gesunken.values()), lambda s: s)
-    assert time.monotonic() - t0 < 1.0  # nicht auf j3 gewartet
-    assert gesunken == {1: "a", 2: "a"}
-
-
-async def test_panel_2of3_dissens_wartet_auf_dritten():
-    """Uneinige erste zwei → der dritte wird abgewartet (Mehrheit braucht ihn)."""
-    import asyncio as aio
-    gesunken = {}
-
-    async def judge(j, delay, antwort):
-        await aio.sleep(delay)
-        return (0, antwort, "")
-
-    tasks = {aio.create_task(judge(1, 0.01, "a")): 1,
-             aio.create_task(judge(2, 0.02, "b")): 2,
-             aio.create_task(judge(3, 0.1, "a")): 3}
-    await blx._panel_2of3(tasks, lambda j, r: gesunken.__setitem__(j, r[1]),
-                          lambda: list(gesunken.values()), lambda s: s)
-    assert gesunken == {1: "a", 2: "b", 3: "a"}
-
-
-async def test_facts_check_inline_cited_evidence(sub_env, monkeypatch, tmp_path):
-    """Facts-Check: zitierte Zeilenbereiche gehen inline mit, Judge läuft ohne Tools,
-    die Check-Datei schreibt die Engine aus der Text-Antwort."""
-    db, ctx, files = sub_env
-    d = _corpus(tmp_path)
-    monkeypatch.setattr(blx, "source_folder", lambda t: d)
-    facts = {"facts": [{"block": "Alpha", "subblock": "Sub Eins", "key_points": ["k"],
-                        "prerequisites": "", "hurdles": "",
-                        "cited_facts": [{"text": "T", "source": "Skript.txt, Z.2"}],
-                        "example_idea": ""}]}
-
-    sh = blx._subs_hash({"Alpha": ["Sub Eins"]})  # Resume-Dateien tragen den Sub-Satz-Hash
-
-    async def fake_slot(ctx2, label, *, key, prompt, role, capabilities, payload, timeout, on_line=None):
-        if "-facts-erg-" in key:
-            return blx.FAILED, None
-        (tmp_path / f"facts-{sh}-c0.json").write_text(json.dumps(facts), encoding="utf-8")
-        return blx.OK, None
-
-    seen = []
-
-    async def fake_agent(key, prompt, timeout, provider="", role="", capabilities="", scope=None, label="", **kw):
-        seen.append((key, capabilities, prompt))
-        return (0, '{"ok": true}', "")
-
-    monkeypatch.setattr(blx, "run_single_slot", fake_slot)
-    monkeypatch.setattr(blx, "run_agent", fake_agent)
-    res = await blx._facts_block(ctx, lambda *a, **k: None, {"arbeit": tmp_path},
-                                 {"Alpha": ["Sub Eins"]}, {"type": "uni"}, d, "", ns="x-")
-    assert res is not None
-    facts_map, discarded = res
-    assert "Alpha" in facts_map and not discarded
-    assert len(seen) == blx.FACTS_CHECK_PANEL
-    key, caps, prompt = seen[0]
-    assert caps == "none" and "── Skript.txt · Z." in prompt
-    assert (tmp_path / f"facts-check-{sh}-c0-j1.json").exists()  # Engine persistiert die Antwort
-
-
-async def test_facts_find_inline_evidence_no_tools(sub_env, monkeypatch, tmp_path):
-    """Facts find/erg mit Korpus: Auszüge inline, Agent ohne Tools — Tool-Agenten
-    verloren sich in Reasoning-Schleifen und endeten mit leerem Turn (Retry-Wellen)."""
-    db, ctx, files = sub_env
-    d = _corpus(tmp_path)
-    monkeypatch.setattr(blx, "source_folder", lambda t: d)
-    seen = []
-    facts = {"facts": [{"block": "Alpha", "subblock": "Sub Eins", "key_points": ["k"],
-                        "prerequisites": "", "hurdles": "",
-                        "cited_facts": [{"text": "T", "source": "Skript.txt, Z.2"}],
-                        "example_idea": ""}]}
-
-    async def fake_slot(ctx2, label, *, key, prompt, role, capabilities, payload, timeout, on_line=None):
-        seen.append((key, capabilities, prompt))
-        return blx.OK, payload((0, json.dumps(facts), ""))
-
-    async def fake_agent(key, prompt, timeout, **kw):  # Check-Panel
-        return (0, '{"ok": true}', "")
-
-    monkeypatch.setattr(blx, "run_single_slot", fake_slot)
-    monkeypatch.setattr(blx, "run_agent", fake_agent)
-    res = await blx._facts_block(ctx, lambda *a, **k: None, {"arbeit": tmp_path},
-                                 {"Alpha": ["Sub Eins"]}, {"type": "uni"}, d, "", ns="x-")
-    assert res is not None
-    finder = [s for s in seen if "-facts-c0" in s[0] or "-facts-erg-" in s[0]]
-    assert finder and all(caps == "none" for _, caps, _ in finder)
-    assert all("── Skript.txt" in prompt for _, _, prompt in finder)  # Auszüge inline
-    assert all("ls/find" not in prompt for _, _, prompt in finder)
-
-
 def test_sub_key_resolves_short_titles():
     """Artefakt-Agenten echoen den Kurztitel; der Sub-Key heißt 'kurztitel: beschreibung'.
     Eindeutiger Präfix wird aufgelöst, Mehrdeutiges und Fehlendes bleibt unverändert."""

backend/tests/test_train.py

@@ -65,17 +65,17 @@ def _trainer(tmp_path, runner, f0, **kw):
 
 
 async def test_aco_konvergiert_auf_optimum(tmp_path):
-    """Gepflanztes Optimum (FACTS_CHUNK_SUBS=6) wird gefunden und bestätigt übernommen;
+    """Gepflanztes Optimum (GEN_PANEL=3) wird gefunden und bestätigt übernommen;
     die Pheromon-Spur konzentriert sich dort."""
     def bewertung(params):
-        return _metrics(note=9.5, dauer=7.0) if params.get("FACTS_CHUNK_SUBS") == 6 else _metrics()
+        return _metrics(note=9.5, dauer=7.0) if params.get("GEN_PANEL") == 3 else _metrics()
 
     runner, f0 = _stub(bewertung)
     t = _trainer(tmp_path, runner, f0, max_trials=120)
     best = await t.run()
-    assert best.get("FACTS_CHUNK_SUBS") == 6
-    taus = t.pheromon["FACTS_CHUNK_SUBS"]
-    assert max(taus, key=lambda k: taus[k]) == "6"
+    assert best.get("GEN_PANEL") == 3
+    taus = t.pheromon["GEN_PANEL"]
+    assert max(taus, key=lambda k: taus[k]) == "3"
 
 
 async def test_uebernahme_braucht_bestaetigung(tmp_path):
@@ -83,7 +83,7 @@ async def test_uebernahme_braucht_bestaetigung(tmp_path):
     zustand = {"mal": 0}
 
     def bewertung(params):
-        if params.get("FACTS_CHUNK_SUBS") == 6:
+        if params.get("GEN_PANEL") == 3:
             zustand["mal"] += 1
             return _metrics(note=9.5) if zustand["mal"] == 1 else _metrics(note=8.0)
         return _metrics()
@@ -91,7 +91,7 @@ async def test_uebernahme_braucht_bestaetigung(tmp_path):
     runner, f0 = _stub(bewertung)
     t = _trainer(tmp_path, runner, f0, max_trials=40)
     best = await t.run()
-    assert best.get("FACTS_CHUNK_SUBS") != 6
+    assert best.get("GEN_PANEL") != 3
 
 
 async def test_f0_filter_verwirft_kaputte_kandidaten(tmp_path):
@@ -111,16 +111,16 @@ async def test_f0_filter_verwirft_kaputte_kandidaten(tmp_path):
 
 async def test_resume_laedt_pheromon_und_cache(tmp_path):
     def bewertung(params):
-        return _metrics(note=9.5) if params.get("FACTS_CHUNK_SUBS") == 6 else _metrics()
+        return _metrics(note=9.5) if params.get("GEN_PANEL") == 3 else _metrics()
 
     runner, f0 = _stub(bewertung)
     t = _trainer(tmp_path, runner, f0, max_trials=60)
     await t.run()
-    best, tau = t.best_params, dict(t.pheromon["FACTS_CHUNK_SUBS"])
+    best, tau = t.best_params, dict(t.pheromon["GEN_PANEL"])
     runner2, f02 = _stub(bewertung)
     t2 = _trainer(tmp_path, runner2, f02, max_trials=0)  # kein Budget: alles aus Persistenz
     assert t2.best_params == best
-    assert t2.pheromon["FACTS_CHUNK_SUBS"] == tau
+    assert t2.pheromon["GEN_PANEL"] == tau
 
 
 async def test_budget_stoppt(tmp_path):