feat: add guarded production Flow v2 authoritative simulation

app/admin_ui/labels.py

@@ -8,6 +8,10 @@ from __future__ import annotations
 from typing import Any
 
 ACTION_LABELS = {
+    "CREATE_PROFORMA": "Criar proforma",
+    "CREATE_INVOICE": "Criar fatura",
+    "VALIDATE_ODOO_ORDER": "Validar encomenda",
+    "REVIEW_REQUIRED": "Rever processo",
     "CALL_CUSTOMER": "Ligar ao cliente",
     "SEND_INFO": "Enviar informação",
     "SEND_QUOTE": "Preparar orçamento",
@@ -38,6 +42,10 @@ ACTION_LABELS = {
 }
 
 PRIMARY_ACTION_LABELS = {
+    "CREATE_PROFORMA": "Criar proforma",
+    "CREATE_INVOICE": "Criar fatura",
+    "VALIDATE_ODOO_ORDER": "Validar encomenda",
+    "REVIEW_REQUIRED": "Rever processo",
     "CALL_CUSTOMER": "Ligar ao cliente",
     "SEND_INFO": "Preparar resposta",
     "SEND_QUOTE": "Preparar orçamento",

app/authoritative_operational_adapter.py

@@ -0,0 +1,182 @@
+"""Authoritative BLIF Flow v2 -> operator decision adapter.
+
+The persisted projection owns factual business state.  Pending tasks are only
+operational obligations and may override that state when this policy proves
+that they are current.  This module is deliberately pure and performs no I/O.
+"""
+from __future__ import annotations
+
+from dataclasses import asdict, dataclass
+from datetime import datetime, timezone
+from typing import Any, Iterable, Mapping
+
+from app.admin_ui.labels import action_label
+
+
+FORMAL_ACTIONS = {"CREATE_PROFORMA", "CREATE_INVOICE"}
+REVIEW_ACTIONS = {"REVIEW", "REVIEW_MANUALLY", "REVIEW_REQUIRED", "REVIEW_RECONSTRUCTED_PROCESS"}
+PAYMENT_FOLLOWUPS = {"FOLLOW_UP_PAYMENT", "FOLLOW_UP_PROFORMA"}
+VALID_OVERRIDES = REVIEW_ACTIONS | {"SUPPORT", "SEND_INFO", "CALL_CUSTOMER", "FOLLOW_UP_CUSTOMER_REVIEW"} | PAYMENT_FOLLOWUPS
+TERMINAL_STATES = {"COMPLETED", "LOST", "NO_INTEREST"}
+
+
+@dataclass(frozen=True)
+class AuthoritativeOperationalDecision:
+    opportunity_id: str
+    canonical_opportunity_id: str
+    business_state: str
+    business_next_action: str | None
+    effective_action: str | None
+    queue: str
+    eligible: bool
+    reason_code: str
+    reason_text: str
+    blocking_action_code: str | None = None
+    obligation_source_refs: tuple[dict[str, Any], ...] = ()
+    confidence: str = "high"
+
+    def to_dict(self) -> dict[str, Any]:
+        result = asdict(self)
+        result["obligation_source_refs"] = list(self.obligation_source_refs)
+        # Shared legacy presentation contract consumed by Operations and detail.
+        result.update({
+            "action_code": self.effective_action or "NO_ACTION",
+            "label": action_label(self.effective_action, "Sem ação"),
+            "description": self.reason_text,
+            "priority": "alta" if self.queue in {"review", "blocked", "exception"} else "normal",
+            "can_execute": self.eligible and self.effective_action is not None,
+            "reason_if_blocked": self.reason_code if not self.eligible else None,
+            "operational_queue": self.queue,
+            "authoritative_v2": True,
+            "suppress_current_card": self.queue == "not_current",
+        })
+        return result
+
+
+def _code(value: Any) -> str:
+    return str(value or "").strip().upper()
+
+
+def _dt(value: Any) -> datetime | None:
+    if not value:
+        return None
+    if isinstance(value, datetime):
+        parsed = value
+    else:
+        try:
+            parsed = datetime.fromisoformat(str(value).replace("Z", "+00:00"))
+        except ValueError:
+            return None
+    return parsed if parsed.tzinfo else parsed.replace(tzinfo=timezone.utc)
+
+
+def _active_obligations(obligations: Iterable[Mapping[str, Any]]) -> list[Mapping[str, Any]]:
+    return [row for row in obligations
+            if str(row.get("status") or "").lower() == "pending"
+            and not row.get("resolved_at") and not row.get("superseded_by_task_id")]
+
+
+def _ref(row: Mapping[str, Any]) -> dict[str, Any]:
+    return {"source": "task", "id": str(row.get("id") or ""), "status": "pending",
+            "action_code": _code(row.get("action_code"))}
+
+
+def decide_authoritative_operation(
+    projection: Mapping[str, Any] | None,
+    *, obligations: Iterable[Mapping[str, Any]] = (), now: datetime | None = None,
+    fiscal_complete: bool = True, reconciliation_blocking: bool = False,
+    hard_blocker: str | None = None,
+) -> AuthoritativeOperationalDecision:
+    """Return the sole operator decision, failing closed without a projection."""
+    now = now or datetime.now(timezone.utc)
+    if projection is None:
+        return AuthoritativeOperationalDecision(
+            "", "", "MISSING_PROJECTION", None, "REVIEW_REQUIRED", "review", True,
+            "MISSING_V2_PROJECTION", "A projeção Flow v2 está em falta; é necessária revisão, sem recorrer ao V1.",
+            confidence="low",
+        )
+    oid = str(projection.get("opportunity_id") or "")
+    canonical = str(projection.get("canonical_opportunity_id") or oid)
+    state = _code(projection.get("business_state"))
+    business_action = _code(projection.get("business_next_action")) or None
+    confidence = str(projection.get("confidence") or "low")
+    if projection.get("is_duplicate_representation"):
+        return AuthoritativeOperationalDecision(
+            oid, canonical, state, business_action, None, "not_current", False,
+            "DUPLICATE_SUPPRESSED", f"Representação duplicada do processo material canónico {canonical}.", confidence="high",
+        )
+    if hard_blocker:
+        return AuthoritativeOperationalDecision(
+            oid, canonical, state, business_action, hard_blocker, "blocked", True,
+            "HARD_FACTUAL_BLOCKER", "Um bloqueio factual ou de sistema impede a ação atual.", hard_blocker, confidence=confidence,
+        )
+
+    active = _active_obligations(obligations)
+    by_code: dict[str, list[Mapping[str, Any]]] = {}
+    for row in active:
+        by_code.setdefault(_code(row.get("action_code")), []).append(row)
+
+    # A formal-document prerequisite blocks only a transition which needs it.
+    if business_action in FORMAL_ACTIONS and not fiscal_complete:
+        return AuthoritativeOperationalDecision(
+            oid, canonical, state, business_action, "VALIDATE_FISCAL_CUSTOMER", "blocked", True,
+            "FISCAL_IDENTITY_REQUIRED", "Validar os dados fiscais antes de criar o documento oficial.",
+            "VALIDATE_FISCAL_CUSTOMER", confidence=confidence,
+        )
+    if business_action in FORMAL_ACTIONS and reconciliation_blocking:
+        return AuthoritativeOperationalDecision(
+            oid, canonical, state, business_action, "RECONCILE_DOCUMENTS", "blocked", True,
+            "DOCUMENT_RECONCILIATION_REQUIRED", "Confirmar a ligação do documento formal atual.",
+            "RECONCILE_DOCUMENTS", confidence=confidence,
+        )
+
+    def choose(codes: Iterable[str], queue: str, reason: str):
+        for code in codes:
+            rows = by_code.get(code, [])
+            if rows:
+                return AuthoritativeOperationalDecision(
+                    oid, canonical, state, business_action, code, queue, True, reason,
+                    "Existe uma obrigação operacional pendente e válida.",
+                    obligation_source_refs=tuple(_ref(row) for row in rows), confidence=confidence,
+                )
+        return None
+
+    picked = choose(("REVIEW_MANUALLY", "REVIEW_REQUIRED", "REVIEW", "REVIEW_RECONSTRUCTED_PROCESS"), "review", "ACTIVE_REVIEW_OBLIGATION")
+    picked = picked or choose(("SUPPORT",), "do_now", "ACTIVE_SUPPORT_OBLIGATION")
+    picked = picked or choose(("SEND_INFO",), "do_now", "ACTIVE_RESPONSE_OBLIGATION")
+    picked = picked or choose(("CALL_CUSTOMER",), "do_now", "EXPLICIT_CALL_OBLIGATION")
+    if picked:
+        return picked
+
+    waiting_state = state in {"AWAITING_CUSTOMER", "AWAITING_PAYMENT"}
+    followup_order = ("FOLLOW_UP_CUSTOMER_REVIEW",) if state == "AWAITING_CUSTOMER" else tuple(PAYMENT_FOLLOWUPS)
+    if waiting_state:
+        for code in followup_order:
+            rows = by_code.get(code, [])
+            if not rows:
+                continue
+            due_rows = [row for row in rows if _dt(row.get("due_at")) is None or _dt(row.get("due_at")) <= now]
+            if due_rows:
+                return AuthoritativeOperationalDecision(
+                    oid, canonical, state, business_action, code, "do_now", True, "DUE_FOLLOW_UP",
+                    "O follow-up ativo chegou à data e continua por satisfazer.",
+                    obligation_source_refs=tuple(_ref(row) for row in due_rows), confidence=confidence,
+                )
+            return AuthoritativeOperationalDecision(
+                oid, canonical, state, business_action, None, "waiting", False, "FOLLOW_UP_NOT_DUE",
+                "O follow-up ativo ainda não chegou à data.",
+                obligation_source_refs=tuple(_ref(row) for row in rows), confidence=confidence,
+            )
+
+    if state in TERMINAL_STATES:
+        return AuthoritativeOperationalDecision(oid, canonical, state, business_action, None, "not_current", False,
+                                                "TERMINAL_BUSINESS_STATE", "O processo factual está concluído.", confidence=confidence)
+    if business_action:
+        queue = "review" if business_action in REVIEW_ACTIONS else "do_now"
+        return AuthoritativeOperationalDecision(oid, canonical, state, business_action, business_action, queue, True,
+                                                "BUSINESS_NEXT_ACTION", str(projection.get("reason_text") or "A ação decorre do estado factual Flow v2."), confidence=confidence)
+    if waiting_state:
+        return AuthoritativeOperationalDecision(oid, canonical, state, None, None, "waiting", False,
+                                                "WAITING_EXTERNAL_EVENT", "O processo aguarda um evento externo.", confidence=confidence)
+    return AuthoritativeOperationalDecision(oid, canonical, state, None, None, "backlog", False,
+                                            "NO_CURRENT_INTERNAL_ACTION", "Não existe ação interna atual.", confidence=confidence)

app/blif_flow_v2_projection_service.py

@@ -19,6 +19,7 @@ from app.config import settings
 
 FLOW_VERSION = "blif-flow-v2-shadow-3"
 WRITE_DATABASE_ALLOWLIST = frozenset({"clientflow_codex_test"})
+PROJECTION_WRITE_TABLES = frozenset({"opportunity_flow_state_v2", "opportunity_flow_transitions"})
 
 
 def _jsonable(value: Any) -> Any:
@@ -78,16 +79,24 @@ def _projection_value(row: dict[str, Any], derived_at: datetime) -> dict[str, An
     return stable | {"source_fingerprint": fingerprint, "derived_at": derived_at}
 
 
-def _derive_all() -> list[dict[str, Any]]:
+def _derive_all(
+    *,
+    expected_database: str = "clientflow_codex_test",
+    expected_user: str | None = "clientflow_codex_test",
+    expected_opportunity_count: int | None = 328,
+    require_opportunities: bool = False,
+) -> list[dict[str, Any]]:
     # Reuse the validated shadow evidence adapter without making it authoritative.
     from scripts.simulate_blif_flow_v2 import collect
 
     report = collect(
-        expected_database="clientflow_codex_test",
-        expected_user="clientflow_codex_test",
+        expected_database=expected_database,
+        expected_user=expected_user,
         # collect() still opens its factual read phase with BEGIN READ ONLY;
         # the session default may be read-write in the isolated test database.
         require_read_only=False,
+        expected_opportunity_count=expected_opportunity_count,
+        require_opportunities=require_opportunities,
     )
     return list(report["opportunities"])
 
@@ -99,6 +108,10 @@ def rebuild_blif_flow_v2_projection(
     allowed_databases: frozenset[str] = WRITE_DATABASE_ALLOWLIST,
     target_schema: str = "public",
     connection: Any | None = None,
+    derive_expected_database: str = "clientflow_codex_test",
+    derive_expected_user: str | None = "clientflow_codex_test",
+    expected_opportunity_count: int | None = 328,
+    require_opportunities: bool = False,
 ) -> dict[str, Any]:
     """Idempotently rebuild projection rows and state-change transitions.
 
@@ -114,7 +127,14 @@ def rebuild_blif_flow_v2_projection(
     if not re.fullmatch(r"[a-z_][a-z0-9_]*", target_schema):
         raise ValueError("invalid target_schema")
 
-    rows = list(derived_rows) if derived_rows is not None else _derive_all()
+    rows = list(derived_rows) if derived_rows is not None else _derive_all(
+        expected_database=derive_expected_database,
+        expected_user=derive_expected_user,
+        expected_opportunity_count=expected_opportunity_count,
+        require_opportunities=require_opportunities,
+    )
+    if require_opportunities and not rows:
+        raise RuntimeError("Flow v2 projection requires at least one opportunity")
     derived_at = datetime.now(timezone.utc)
     values = [_projection_value(row, derived_at) for row in rows]
     if len({value["opportunity_id"] for value in values}) != len(values):

app/canonical_operations.py

@@ -321,13 +321,14 @@ def _opportunity_item(
     rows: list[Mapping[str, Any]],
     decision: Mapping[str, Any],
 ) -> tuple[dict[str, Any] | None, list[dict[str, Any]]]:
+    authoritative = decision.get("authoritative_v2") is True
     scheduled_call = next((
         row for row in rows
         if row.get("source") == "task"
         and _s(row.get("status")).lower() == "pending"
         and _canonical_code(row.get("action_code")) == "CALL_CUSTOMER"
     ), None)
-    if scheduled_call:
+    if scheduled_call and not authoritative:
         decision = {
             **dict(decision),
             "action_code": "CALL_CUSTOMER",
@@ -347,19 +348,21 @@ def _opportunity_item(
             evidence_rows.append(row)
 
     if action_code in NO_WORK_ACTION_CODES:
+        if decision.get("suppress_current_card"):
+            return None, independent_exceptions
+        keep_authoritative_noncurrent = authoritative and _s(decision.get("operational_queue")) in {"waiting", "backlog"}
+        if not keep_authoritative_noncurrent:
             if not _has_explicit_unresolved_review(evidence_rows):
                 return None, independent_exceptions
-        # The central decision normally wins, but an explicit unresolved review
-        # marker must not be hidden by NO_ACTION/NOT_FOUND.
+            # The central decision normally wins, but an explicit unresolved
+            # review marker must not be hidden by NO_ACTION/NOT_FOUND.
             action_code = "REVIEW"
             decision = {
-            **dict(decision),
-            "action_code": action_code,
+                **dict(decision), "action_code": action_code,
                 "label": "Rever evidência pendente",
                 "description": "A decisão central indica que não há ação, mas existe evidência explicitamente marcada para revisão.",
                 "reason": "Existe revisão humana não resolvida apesar da decisão central sem ação.",
-            "priority": "normal",
-            "target_url": f"/opportunities/{opportunity_id}",
+                "priority": "normal", "target_url": f"/opportunities/{opportunity_id}",
             }
 
     matching_task = scheduled_call or _matching_task(evidence_rows, action_code)
@@ -367,6 +370,7 @@ def _opportunity_item(
     primary = primary or (evidence_rows[0] if evidence_rows else {})
     process_key = f"opportunity:{opportunity_id}"
     waiting = _is_waiting(decision)
+    decision_queue = _s(decision.get("operational_queue")) if authoritative else ""
     opportunity_row = next((row for row in evidence_rows if _s(row.get("fiscal_customer_name"))), None)
     opportunity_row = opportunity_row or next((row for row in evidence_rows if _s(row.get("opportunity_title"))), None)
     opportunity_row = opportunity_row or primary
@@ -383,7 +387,7 @@ def _opportunity_item(
         "detail": _s(decision.get("description") or decision.get("reason")),
         "why_human_required": _s(decision.get("reason") or decision.get("description")),
         "priority": _s(decision.get("priority")) or _best_priority(evidence_rows),
-        "operational_queue": "waiting" if waiting else "do_now",
+        "operational_queue": decision_queue or ("waiting" if waiting else "do_now"),
         "queue": _s((matching_task or primary).get("queue")) or _s(primary.get("queue")) or "rever",
         "status": "waiting" if waiting else "pending",
         "href": _s((matching_task or {}).get("href")) or _s(decision.get("target_url")) or f"/opportunities/{opportunity_id}",
@@ -400,6 +404,10 @@ def _opportunity_item(
         ],
         "decision_version": decision.get("decision_version"),
         "decision": dict(decision),
+        "business_state": decision.get("business_state"),
+        "business_next_action": decision.get("business_next_action"),
+        "effective_action": decision.get("effective_action"),
+        "canonical_opportunity_id": decision.get("canonical_opportunity_id") or opportunity_id,
     })
     return item, independent_exceptions
 

app/operational_eligibility.py

@@ -157,6 +157,23 @@ def apply_operational_eligibility(items: Iterable[Mapping[str, Any]], *, now: da
     projected = []
     for source in items:
         item = dict(source)
+        decision = item.get("decision") if isinstance(item.get("decision"), Mapping) else {}
+        if decision.get("authoritative_v2") is True:
+            item["eligibility"] = {
+                "eligible": bool(decision.get("eligible")),
+                "queue": decision.get("operational_queue"),
+                "reason_code": decision.get("reason_code"),
+                "reason_text": decision.get("description"),
+                "blocking_action_code": decision.get("blocking_action_code"),
+                "obligation_source_refs": decision.get("obligation_source_refs") or [],
+                "confidence": decision.get("confidence"),
+            }
+            item["operational_queue"] = decision.get("operational_queue")
+            item["why_human_required"] = decision.get("description")
+            item["eligibility_reason_code"] = decision.get("reason_code")
+            item["eligible"] = bool(decision.get("eligible"))
+            projected.append(item)
+            continue
         eligibility = evaluate_operational_eligibility(item, now=now)
         item["eligibility"] = eligibility.to_dict()
         item["operational_queue"] = eligibility.queue

app/operations_service.py

@@ -8,6 +8,7 @@ from __future__ import annotations
 import os
 import re
 import subprocess
+from contextlib import nullcontext
 from typing import Any, Dict, List, Optional
 
 from sqlalchemy import text
@@ -26,7 +27,7 @@ from app.canonical_operations import (
     opportunity_ids_from_work_seeds,
     partition_canonical_items,
 )
-from app.opportunity_next_action_service import get_opportunity_next_actions
+from app.opportunity_next_action_service import get_opportunity_next_actions, get_opportunity_next_actions_for_mode
 from app.document_reconciliation_service import active_document_link_exclusion_sql
 
 
@@ -104,7 +105,7 @@ def _norm_identity(value: Any) -> str:
 
 
 def _identity_tokens(value: Any) -> set[str]:
-    return {
+    result = {
         token
         for token in _norm_identity(value).split()
         if len(token) >= 3 and token not in _IDENTITY_STOPWORDS
@@ -302,7 +303,9 @@ def _attach_operation_urls(items: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
     return _sanitize_operation_identities(cleaned)
 
 
-def get_operations_summary(limit: int = 24) -> Dict[str, Any]:
+def get_operations_summary(
+    limit: int = 24, *, flow_mode: str | None = None, connection: Any = None,
+) -> Dict[str, Any]:
     """Build the /operations work queue summary.
 
     /operations is intentionally not a mini-dashboard. It returns a compact
@@ -315,7 +318,8 @@ def get_operations_summary(limit: int = 24) -> Dict[str, Any]:
     # after canonicalization. This prevents duplicate rows from consuming the
     # display limit while still avoiding an unbounded history scan.
     candidate_limit = max(200, min(display_limit * 20, 1000))
-    with engine.begin() as conn:
+    active_flow_mode = str(flow_mode if flow_mode is not None else settings.blif_flow_v2_mode or "off").strip().lower()
+    with (nullcontext(connection) if connection is not None else engine.begin()) as conn:
         counts = conn.execute(text("""
             SELECT
               (SELECT COUNT(*) FROM opportunities WHERE status = 'open')::int AS open_opportunities,
@@ -672,12 +676,50 @@ def get_operations_summary(limit: int = 24) -> Dict[str, Any]:
             conversation_timeline = {str(row["conversation_id"]): dict(row) for row in timeline_rows}
 
     seed_rows = [dict(r) for r in work_seed_rows]
+    # In authoritative mode the factual projection, not a legacy task/stage,
+    # selects material processes that have current business work.  Synthetic
+    # rows are read-model seeds only and are never persisted.
+    if active_flow_mode == "authoritative":
+        seeded = {str(row.get("opportunity_id") or "") for row in seed_rows}
+        with (nullcontext(connection) if connection is not None else engine.connect()) as conn:
+            v2_seeds = conn.execute(text("""
+                SELECT p.opportunity_id::text, p.business_next_action, p.derived_at,
+                       o.title, o.value_amount, o.currency, o.lifecycle_state,
+                       c.name AS customer_name
+                FROM opportunity_flow_state_v2 p
+                JOIN opportunities o ON o.id = p.opportunity_id
+                LEFT JOIN customers c ON c.id = o.local_customer_id
+                WHERE p.is_duplicate_representation = false
+                  AND (p.business_next_action IS NOT NULL
+                       OR p.business_state IN ('AWAITING_CUSTOMER', 'AWAITING_PAYMENT'))
+            """)).mappings().all()
+        for projection_seed in v2_seeds:
+            oid = str(projection_seed["opportunity_id"])
+            if oid in seeded:
+                continue
+            seed_rows.append({
+                "source": "flow_v2", "id": oid, "opportunity_id": oid,
+                "created_at": projection_seed.get("derived_at"), "due_at": None,
+                "priority": "normal", "queue": "vendas",
+                "action_code": projection_seed.get("business_next_action"), "status": "projected",
+                "title": projection_seed.get("title") or "Oportunidade",
+                "opportunity_title": projection_seed.get("title") or "",
+                "customer_name": projection_seed.get("customer_name") or "",
+                "fiscal_customer_name": projection_seed.get("customer_name") or "",
+                "item_metadata": {}, "opportunity_metadata": {},
+                "opportunity_lifecycle_state": projection_seed.get("lifecycle_state") or "",
+                "opportunity_value_amount": projection_seed.get("value_amount") or 0,
+                "opportunity_currency": projection_seed.get("currency") or "EUR",
+                "href": f"/opportunities/{oid}", "action_label": "Abrir",
+            })
     for row in seed_rows:
         timeline = conversation_timeline.get(str(row.get("conversation_id") or ""), {})
         row.update({key: timeline.get(key) for key in ("latest_public_inbound", "latest_public_outbound")})
     normalized_rows = _normalise_work_item_intent(seed_rows)
     opportunity_ids = opportunity_ids_from_work_seeds(normalized_rows)
-    decisions = get_opportunity_next_actions(opportunity_ids) if opportunity_ids else {}
+    decisions = (get_opportunity_next_actions_for_mode(
+        opportunity_ids, flow_mode=active_flow_mode, connection=connection,
+    ) if opportunity_ids else {})
     projection = canonicalize_operations(normalized_rows, decisions, evidence_rows=[dict(row) for row in evidence_rows])
     canonical_items = _attach_operation_urls(list(projection["items"]))
     partition = partition_canonical_items(canonical_items, display_limit=display_limit)
@@ -722,6 +764,11 @@ def get_operations_summary(limit: int = 24) -> Dict[str, Any]:
         **diagnostics,
         "projection_metrics": diagnostics,
     }
+    if flow_mode is not None:
+        # Complete, untruncated read-model population for offline simulation;
+        # normal API/GET callers do not receive this audit-only field.
+        result["simulation_all_items"] = canonical_items
+    return result
 
 
 def get_system_health_summary() -> Dict[str, Any]:

app/opportunity_next_action_service.py

@@ -15,6 +15,7 @@ from sqlalchemy import bindparam, text
 
 from app.db import engine
 from app.config import settings
+from app.authoritative_operational_adapter import decide_authoritative_operation
 # Backward-compatible static anchors from v1.5.59: quotation_doc, invoice_doc, confirmar pagamento antes de emitir fatura, Pagamento confirmado com base em, Criar/enviar fatura.
 from app.domain.opportunity_flow import (
     OpportunityEvidence,
@@ -51,6 +52,11 @@ def _load_v2_comparison_rows(opportunity_ids: list[str]) -> dict[str, dict[str,
     """Read the persisted projection only; comparison must never derive writes."""
     if not opportunity_ids:
         return {}
+    # Lightweight repository fakes used by pure unit tests intentionally expose
+    # only ``begin``.  Comparison observation is optional and must not alter the
+    # V1 contract or its query count when that read capability is absent.
+    if not hasattr(engine, "connect"):
+        return {}
     with engine.connect() as conn:
         rows = conn.execute(text("""
             SELECT opportunity_id::text, material_process_key,
@@ -95,8 +101,6 @@ def compare_v1_v2_decisions(
 
 def _observe_flow_v2(v1_decisions: Dict[str, Dict[str, Any]]) -> None:
     mode = _flow_v2_mode()
-    if mode == "authoritative":
-        raise RuntimeError("BLIF Flow v2 authoritative mode is disabled; cutover boundary is fail-closed")
     if mode != "compare" or not v1_decisions:
         return
     rows = _load_v2_comparison_rows(list(v1_decisions))
@@ -228,7 +232,7 @@ def get_opportunity_next_action(
     """
 
     if _flow_v2_mode() == "authoritative":
-        raise RuntimeError("BLIF Flow v2 authoritative mode is disabled; cutover boundary is fail-closed")
+        return get_opportunity_next_actions([opportunity_id]).get(opportunity_id) or decide_authoritative_operation(None).to_dict()
     evidence = (_build_db_evidence(opportunity_id, preloaded=preloaded)
                 if preloaded is not None else _build_db_evidence(opportunity_id))
     if evidence is None:
@@ -368,16 +372,28 @@ def _bulk_operation_snapshots(
 
 def get_opportunity_next_actions(opportunity_ids: Iterable[str]) -> Dict[str, Dict[str, Any]]:
     """Return the same decisions as the single-item API with a fixed query count."""
-    if _flow_v2_mode() == "authoritative":
-        raise RuntimeError("BLIF Flow v2 authoritative mode is disabled; cutover boundary is fail-closed")
     ids = list(dict.fromkeys(str(value).strip() for value in opportunity_ids if str(value).strip()))
     if not ids:
         return {}
 
+    return get_opportunity_next_actions_for_mode(ids, flow_mode=_flow_v2_mode())
+
+
+def get_opportunity_next_actions_for_mode(
+    opportunity_ids: Iterable[str], *, flow_mode: str, connection: Any = None,
+) -> Dict[str, Dict[str, Any]]:
+    """Explicit-mode read boundary used by the read-only cutover simulation."""
+    ids = list(dict.fromkeys(str(value).strip() for value in opportunity_ids if str(value).strip()))
+    if not ids:
+        return {}
+    if flow_mode == "authoritative":
+        return _get_authoritative_opportunity_decisions(ids, connection=connection)
+
     # Read path only: schema creation belongs to startup/migrations.  In
     # particular, GET /operations must never perform DDL while calculating its
     # canonical projection.
-    with engine.begin() as conn:
+    from contextlib import nullcontext
+    with (nullcontext(connection) if connection is not None else engine.begin()) as conn:
         opportunities = _bulk_rows(conn, """
             SELECT id::text, stage, status, title,
                    local_customer_id::text AS fiscal_customer_id,
@@ -448,5 +464,60 @@ def get_opportunity_next_actions(opportunity_ids: Iterable[str]) -> Dict[str, Di
             company_profile="blif",
         )
         decisions[oid] = decide_opportunity_next_action(evidence, profile).to_dict()
+    if flow_mode == "compare":
         _observe_flow_v2(decisions)
     return decisions
+
+
+def _get_authoritative_opportunity_decisions(ids: list[str], *, connection: Any = None) -> Dict[str, Dict[str, Any]]:
+    """Read-only set-oriented adapter input loader; never derives or persists V2."""
+    from contextlib import nullcontext
+    with (nullcontext(connection) if connection is not None else engine.connect()) as conn:
+        projections = _bulk_rows(conn, """
+            SELECT opportunity_id::text, material_process_key,
+                   canonical_opportunity_id::text, is_duplicate_representation,
+                   business_state, business_next_action, diagnostic_status,
+                   confidence, reason_code, reason_text
+            FROM opportunity_flow_state_v2 WHERE opportunity_id::text IN :opportunity_ids
+        """, ids)
+        tasks = _bulk_rows(conn, """
+            SELECT id::text, opportunity_id::text, action_code, status, due_at,
+                   resolved_at, superseded_by_task_id::text, created_at, metadata
+            FROM tasks WHERE opportunity_id::text IN :opportunity_ids AND status = 'pending'
+        """, ids)
+        customers = _bulk_rows(conn, """
+            SELECT o.id::text AS opportunity_id,
+                   (c.tax_id IS NOT NULL AND c.tax_id <> '' AND c.email IS NOT NULL AND c.email <> ''
+                    AND c.street_name IS NOT NULL AND c.street_name <> ''
+                    AND c.postal_zone IS NOT NULL AND c.postal_zone <> ''
+                    AND c.city_name IS NOT NULL AND c.city_name <> '') AS fiscal_complete
+            FROM opportunities o LEFT JOIN customers c ON c.id = o.local_customer_id
+            WHERE o.id::text IN :opportunity_ids
+        """, ids)
+        reconciliation = _bulk_rows(conn, """
+            SELECT opportunity_id::text
+            FROM reconciliation_items
+            WHERE opportunity_id::text IN :opportunity_ids
+              AND status IN ('needs_review','conflict')
+              AND COALESCE((payload->>'reconciliation_blocks_current_action')::boolean, false)
+        """, ids)
+    projection_by_id = {str(row["opportunity_id"]): row for row in projections}
+    tasks_by_id = _group_by_opportunity(tasks)
+    fiscal_by_id = {str(row["opportunity_id"]): bool(row.get("fiscal_complete")) for row in customers}
+    reconciliation_ids = {str(row["opportunity_id"]) for row in reconciliation}
+    decisions = {}
+    for oid in ids:
+        projection = projection_by_id.get(oid)
+        decision = decide_authoritative_operation(
+            projection, obligations=tasks_by_id.get(oid, []),
+            fiscal_complete=fiscal_by_id.get(oid, False),
+            reconciliation_blocking=oid in reconciliation_ids,
+        )
+        value = decision.to_dict()
+        if projection is None:
+            value["opportunity_id"] = oid
+            value["target_url"] = f"/opportunities/{oid}"
+        else:
+            value["target_url"] = f"/opportunities/{decision.canonical_opportunity_id or oid}"
+        decisions[oid] = value
+    return decisions

scripts/audit_blif_flow_v2_cutover.py

@@ -1,228 +1,328 @@
 #!/usr/bin/env python3
-"""Read-only compatibility/cutover audit for BLIF Flow v2."""
+"""Run a strictly read-only BLIF Flow v2 shadow/cutover audit."""
 from __future__ import annotations
 
+import argparse
 import json
+import os
 import sys
 from collections import Counter
 from datetime import datetime, timezone
 from pathlib import Path
-from typing import Any
+from typing import Any, Sequence
+
 
 ROOT = Path(__file__).resolve().parents[1]
 sys.path.insert(0, str(ROOT))
+os.chdir(ROOT)
 
-from sqlalchemy import text
+PRODUCTION_DATABASE = "clientflow"
+TEST_DATABASE = "clientflow_codex_test"
+AUDIT = Path("/tmp/blif_flow_v2_production_shadow_audit.json")
+SEMANTIC = Path("/tmp/blif_flow_v2_production_semantic_compare.json")
+SUMMARY = Path("/tmp/blif_flow_v2_production_shadow_summary.txt")
+CLASSIFICATIONS = {
+    "SEMANTICALLY_EQUIVALENT", "V1_OPERATIONAL_OVERRIDE", "V2_CORRECTS_V1",
+    "LEGACY_ONLY", "REAL_CONFLICT", "MISSING_PROJECTION",
+}
+CURRENT_QUEUES = {"do_now", "review", "exception"}
+OVERRIDE_PRECEDENCE = {
+    "scheduled_call", "due_followup", "future_followup", "integration_exception",
+    "document_prerequisite", "fiscal_prerequisite", "safe_preserve_v1",
+}
+LEGACY_ACTIONS = {
+    "CREATE_JASMIN_QUOTE", "NO_ACTION", "WAIT_CUSTOMER", "WAIT_PAYMENT",
+    "WAIT_PRODUCTION", "WAIT_LOGISTICS", "WAIT_SUPPLIER", "WAIT_SCHEDULED_DATE",
+}
+REVIEW_ACTIONS = {
+    "REVIEW", "REVIEW_REQUIRED", "REVIEW_MANUALLY", "REVIEW_RECONSTRUCTED_PROCESS",
+    "RECONCILE_DOCUMENTS", "VALIDATE_FISCAL_CUSTOMER", "REVIEW_EXCEPTION",
+}
+NAMED_IDS = {
+    "INSTALBEIRA": "5c33db95-fab8-477a-bddd-0b9cc8f91302",
+    "PANORAMIC": "fd79b9a1-07e6-4f61-95e8-09eab89c155e",
+    "ENGEXICON": "61f1c955-a372-4ea7-b9b0-b8528d74a141",
+    "CONSTRURECUP": "e3b23ac5-84db-4763-8a31-a684e873032c",
+    "X_MAT_CANONICAL": "dc89a466-db24-401b-bfe9-d47644b2d0c8",
+    "X_MAT_DUPLICATE": "1816a06e-9a69-4a9b-9279-1263156892d3",
+    "RZSOLAR_CANONICAL": "fd221608-e007-4043-a23d-07e0c119a345",
+    "RZSOLAR_DUPLICATE": "434124fb-ac19-4d78-909a-55761d7e8daa",
+}
 
-from app.db import engine
-from app.opportunity_next_action_service import (
-    _load_v2_comparison_rows, compare_v1_v2_decisions, get_opportunity_next_actions,
+
+def build_parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument(
+        "--production-readonly-audit", action="store_true",
+        help="explicitly authorize a read-only audit of database clientflow",
     )
-from scripts.simulate_blif_flow_v2 import collect
+    return parser
 
 
-EXPECTED_DATABASE = "clientflow_codex_test"
-CUTOVER = Path("/tmp/blif_flow_v2_cutover_report.json")
-INVENTORY = Path("/tmp/blif_flow_v2_legacy_field_inventory.json")
-COMPARE = Path("/tmp/blif_flow_v2_compare_report.json")
+def validate_execution(
+    *, production_readonly_audit: bool, database: str, transaction_read_only: str,
+    mode: str,
+) -> None:
+    normalized_mode = str(mode or "").strip().lower()
+    if production_readonly_audit:
+        if database != PRODUCTION_DATABASE:
+            raise RuntimeError(
+                f"--production-readonly-audit requires database {PRODUCTION_DATABASE!r}, found {database!r}"
+            )
+        if transaction_read_only != "on":
+            raise RuntimeError("production audit requires transaction_read_only=on")
+        if normalized_mode not in {"shadow", "compare"}:
+            raise RuntimeError("production audit requires BLIF_FLOW_V2_MODE=shadow or compare")
+        return
+    if database == PRODUCTION_DATABASE:
+        raise RuntimeError("production database requires explicit --production-readonly-audit opt-in")
+    if database != TEST_DATABASE:
+        raise RuntimeError(f"default audit requires database {TEST_DATABASE!r}, found {database!r}")
+    if transaction_read_only != "on":
+        raise RuntimeError("cutover audit requires an explicit READ ONLY transaction")
 
 
-FIELD_INVENTORY = [
-    {
-        "field": "stage", "kind": "derived_compatibility", "factual": False,
-        "writers": ["app/opportunity_service.py", "app/admin_ui/pages/opportunities.py",
-                    "app/reconciliation_service.py", "app/odoo_service.py", "app/external_reconciliation_sync.py"],
-        "readers": ["app/domain/opportunity_flow/evidence.py (V1 only)", "app/workflow_guard.py",
-                    "app/admin_ui/pages/opportunities.py", "app/admin_ui/pages/orders.py",
-                    "app/admin_dashboard.py", "app/revenue_forecast_service.py"],
-        "flow_v2_replacement": "opportunity_flow_state_v2.business_state",
-        "compatibility_required": True, "one_time_migration_required": False,
-        "eventually_deprecatable": True,
-        "notes": "Keep stored value initially for V1 boards, reports, forecasts and action guards; never use as factual V2 input. Consumers need code migration, not historical mass rewrite.",
-    },
-    {
-        "field": "lifecycle_state", "kind": "operational_compatibility", "factual": False,
-        "writers": ["app/followup_service.py", "app/opportunity_service.py"],
-        "readers": ["app/operational_eligibility.py", "app/operations_service.py",
-                    "app/admin_ui/pages/opportunities.py"],
-        "flow_v2_replacement": "business state plus active operational obligations",
-        "compatibility_required": True, "one_time_migration_required": False,
-        "eventually_deprecatable": True,
-        "notes": "Awaiting/recovery/nurture presentation remains V1 compatibility. It cannot create a scheduled obligation without an active task.",
-    },
-    {
-        "field": "next_follow_up_at", "kind": "denormalized_followup_compatibility", "factual": False,
-        "writers": ["app/followup_service.py", "app/opportunity_service.py"],
-        "readers": ["app/operations_service.py", "app/admin_ui/pages/opportunities.py"],
-        "flow_v2_replacement": "pending follow-up task action_code/due_at",
-        "compatibility_required": True, "one_time_migration_required": False,
-        "eventually_deprecatable": True,
-        "notes": "Historical timestamp is inert without a pending follow-up task; retain all 44 values for now.",
-    },
-    {
-        "field": "nurture_until", "kind": "denormalized_schedule_compatibility", "factual": False,
-        "writers": ["app/opportunity_service.py"],
-        "readers": ["app/admin_ui/pages/opportunities.py"],
-        "flow_v2_replacement": "explicit pending REVIEW_NURTURE/follow-up task due_at",
-        "compatibility_required": True, "one_time_migration_required": False,
-        "eventually_deprecatable": True,
-    },
-    {
-        "field": "follow_up_attempts", "kind": "historical_counter", "factual": False,
-        "writers": ["app/opportunity_service.py", "app/followup_service.py"],
-        "readers": ["app/admin_ui/pages/opportunities.py"],
-        "flow_v2_replacement": "event/task history aggregation", "compatibility_required": True,
-        "one_time_migration_required": False, "eventually_deprecatable": True,
-    },
-    {
-        "field": "last_action_code", "kind": "last-known-action_compatibility", "factual": False,
-        "writers": ["app/opportunity_service.py", "app/admin_ui/pages/opportunities.py",
-                    "app/reconciliation_service.py", "app/odoo_service.py"],
-        "readers": ["app/workflow_guard.py", "app/admin_dashboard.py",
-                    "app/admin_ui/pages/opportunities.py", "app/action_prompt.py"],
-        "flow_v2_replacement": "opportunity_flow_state_v2.business_next_action plus OperationalEligibility",
-        "compatibility_required": True, "one_time_migration_required": False,
-        "eventually_deprecatable": True,
-    },
-    {
-        "field": "last_task_id", "kind": "historical_pointer", "factual": False,
-        "writers": ["app/opportunity_service.py", "app/reconciliation_service.py",
-                    "app/company_opportunity_linking.py"],
-        "readers": ["app/workflow_guard.py"],
-        "flow_v2_replacement": "active canonical task lookup", "compatibility_required": True,
-        "one_time_migration_required": False, "eventually_deprecatable": True,
-    },
-    {
-        "field": "pending_primary_* / pending_follow_up_*", "kind": "runtime_read_model", "factual": False,
-        "writers": ["none (SQL projections in app/opportunity_service.py)"],
-        "readers": ["app/admin_ui/pages/opportunities.py"],
-        "flow_v2_replacement": "active task lookup remains an explicit operational overlay",
-        "compatibility_required": True, "one_time_migration_required": False,
-        "eventually_deprecatable": False,
-        "notes": "These virtual fields correctly derive active obligations and are not stored opportunity state.",
-    },
-]
+def _code(value: Any) -> str:
+    return str(value or "").strip().upper()
 
-SOURCE_OF_TRUTH = [
-    ["business process state", "opportunity_flow_state_v2.business_state"],
-    ["business next action", "opportunity_flow_state_v2.business_next_action"],
-    ["time-sensitive operational queue", "OperationalEligibility"],
-    ["scheduled follow-up", "pending task action_code + due_at"],
-    ["formal document state", "commercial_documents + active opportunity_document_links"],
-    ["payment", "confirmed factual payment evidence/operation link"],
-    ["Odoo execution", "operation_links / factual Odoo evidence"],
-    ["messages/customer response", "messages and communications chronology"],
-    ["legacy stage", "compatibility only"],
-    ["tasks", "operator obligations/history; never business fact proof"],
-]
+
+def classify_semantic_difference(
+    *, v1_state: str | None, v1_action: str | None,
+    v2_state: str | None, v2_action: str | None,
+    v2_projection_present: bool = True,
+    is_duplicate_representation: bool = False,
+    operational_action: str | None = None,
+    operational_precedence: str | None = None,
+    operational_queue: str | None = None,
+    v2_confidence: str | None = None,
+    v2_diagnostic_status: str | None = None,
+) -> tuple[str, str]:
+    """Conservatively compare meanings rather than raw action vocabulary."""
+    if not v2_projection_present:
+        return "MISSING_PROJECTION", "No persisted Flow v2 projection exists for this opportunity."
+    v1, v2, effective = _code(v1_action), _code(v2_action), _code(operational_action)
+    precedence = str(operational_precedence or "").strip().lower()
+    queue = str(operational_queue or "").strip().lower()
+    if is_duplicate_representation:
+        return "V2_CORRECTS_V1", "Material identity suppresses a duplicate representation without deleting evidence."
+    if precedence in OVERRIDE_PRECEDENCE and effective and (v1 == effective or queue in CURRENT_QUEUES | {"waiting"}):
+        return "V1_OPERATIONAL_OVERRIDE", f"Explicit operational precedence {precedence} validly overlays the V2 business transition."
+    if v1 == v2 and v1:
+        return "SEMANTICALLY_EQUIVALENT", "V1 and V2 select the same action."
+    if not v1 and not v2:
+        return "SEMANTICALLY_EQUIVALENT", "Neither model has a current business action."
+    if v1 in REVIEW_ACTIONS and (v2 in REVIEW_ACTIONS or _code(v2_state) in {"REVIEW_REQUIRED", "FISCAL_BLOCKED", "DOCUMENT_RECONCILIATION_REQUIRED"}):
+        return "SEMANTICALLY_EQUIVALENT", "Both decisions require review/blocker handling."
+    if v1 in {"NO_ACTION", ""} and not v2 and _code(v2_state) in {"COMPLETED", "LOST", "NO_INTEREST"}:
+        return "SEMANTICALLY_EQUIVALENT", "Both decisions represent a terminal/non-current process."
+    if v1.startswith("FOLLOW_UP_") and _code(v2_state) in {"AWAITING_CUSTOMER", "AWAITING_PAYMENT"}:
+        return "V1_OPERATIONAL_OVERRIDE", "A current follow-up obligation overlays a waiting V2 business state."
+    if v1 == "CONFIRM_PAYMENT" and _code(v2_state) == "AWAITING_PAYMENT" and not v2:
+        return "SEMANTICALLY_EQUIVALENT", "Both decisions mean payment remains outstanding; V1 names the compatibility action."
+    if v1 in {"VALIDATE_FISCAL_CUSTOMER", "RECONCILE_DOCUMENTS"} and not effective and (
+        queue in {"not_current", "backlog"} or _code(v2_state) in {"INQUIRY", "AWAITING_CUSTOMER", "AWAITING_PAYMENT"}
+    ):
+        return "LEGACY_ONLY", "Legacy data-hygiene/blocker vocabulary is not a current factual V2 obligation."
+    if precedence in {"safe_diagnostic_only", "safe_ambiguous_review"}:
+        return "V2_CORRECTS_V1", "SAFE V2 prevents ambiguous historical compatibility state from creating current work."
+    if _code(v2_state) in {"REVIEW_REQUIRED", "FISCAL_BLOCKED", "DOCUMENT_RECONCILIATION_REQUIRED"}:
+        return "V2_CORRECTS_V1", "V2 converts conflicting factual history into an explicit protected review state."
+    if effective and effective == v2 and v1 != v2:
+        return "V2_CORRECTS_V1", "The safe operational action follows the factual V2 transition rather than the legacy action."
+    if v1 in LEGACY_ACTIONS:
+        if v2 and _code(v2_state) not in {"AWAITING_CUSTOMER", "AWAITING_PAYMENT"}:
+            return "V2_CORRECTS_V1", "V2 replaces a legacy compatibility action with a factual business transition."
+        return "LEGACY_ONLY", "V1 action is compatibility vocabulary with no native V2 business transition."
+    if v2 and str(v2_confidence or "").lower() == "high" and str(v2_diagnostic_status or "").lower() == "clear":
+        return "V2_CORRECTS_V1", "High-confidence factual V2 transition corrects a different legacy action."
+    if effective and v1 == effective:
+        return "V1_OPERATIONAL_OVERRIDE", "V1 matches the safe operational overlay rather than the business action."
+    return "REAL_CONFLICT", "The available evidence does not establish equivalence, a valid override, or a safe V2 correction."
 
 
 def _write(path: Path, value: Any) -> None:
     path.write_text(json.dumps(value, ensure_ascii=False, indent=2, default=str), encoding="utf-8")
 
 
-def _identity_and_ids() -> tuple[dict[str, str], list[str], int]:
+def _queue_totals(rows: list[dict[str, Any]], key: str) -> dict[str, int]:
+    counts = Counter(str(row[key].get("operational_queue") or row[key].get("effective_operational_queue") or "not_current") for row in rows)
+    return {
+        "current": sum(counts[name] for name in CURRENT_QUEUES),
+        "do_now": counts["do_now"], "review": counts["review"],
+        "waiting": counts["waiting"], "backlog": counts["backlog"],
+        "not_current": counts["not_current"],
+    }
+
+
+def _operation_totals(source: dict[str, Any]) -> dict[str, int]:
+    return {"current": int(source.get("current_work") or 0),
+            "do_now": int(source.get("do_now") or 0),
+            "review": int(source.get("review") or 0),
+            "waiting": int(source.get("waiting") or 0),
+            "backlog": int(source.get("backlog") or 0),
+            "not_current": int(source.get("not_current") or 0)}
+
+
+def _database_snapshot(engine: Any) -> tuple[dict[str, str], list[str], dict[str, dict[str, Any]]]:
+    from sqlalchemy import text
     with engine.connect() as conn:
         conn.exec_driver_sql("BEGIN READ ONLY")
         try:
-            identity = conn.execute(text("SELECT current_database(),current_user,current_setting('transaction_read_only')")).one()
-            if identity[0] != EXPECTED_DATABASE or identity[2] != "on":
-                raise RuntimeError(f"refusing unexpected/non-read-only database: {identity!r}")
-            ids = [str(value) for value in conn.execute(text("SELECT opportunity_id FROM opportunity_flow_state_v2 ORDER BY opportunity_id")).scalars()]
-            historical = conn.execute(text("""
-                SELECT count(*) FROM opportunities o
-                WHERE o.next_follow_up_at IS NOT NULL
-                  AND NOT EXISTS (
-                    SELECT 1 FROM tasks t WHERE t.opportunity_id=o.id AND t.status='pending'
-                      AND (t.action_code LIKE 'FOLLOW_UP_%' OR t.action_code IN
-                           ('CALL_CUSTOMER','CONFIRM_DELIVERY','RECOVER_OPPORTUNITY','REVIEW_NURTURE'))
-                  )
-            """)).scalar_one()
+            identity = conn.execute(text(
+                "SELECT current_database(),current_user,current_setting('transaction_read_only')"
+            )).one()
+            opportunity_ids = [str(value) for value in conn.execute(text(
+                "SELECT id FROM opportunities ORDER BY id"
+            )).scalars()]
+            rows = conn.execute(text("""
+                SELECT opportunity_id::text, material_process_key,
+                       canonical_opportunity_id::text, is_duplicate_representation,
+                       business_state, business_next_action, diagnostic_status,
+                       confidence, reason_code, reason_text, evidence_refs, flow_version
+                FROM opportunity_flow_state_v2 ORDER BY opportunity_id
+            """)).mappings().all()
         finally:
             conn.rollback()
-    return {"database": identity[0], "user": identity[1], "transaction_read_only": identity[2]}, ids, int(historical)
+    return (
+        {"database": identity[0], "user": identity[1], "transaction_read_only": identity[2]},
+        opportunity_ids, {str(row["opportunity_id"]): dict(row) for row in rows},
+    )
 
 
-def main() -> None:
-    identity, ids, historical_timestamps = _identity_and_ids()
-    if len(ids) != 328:
-        raise RuntimeError(f"expected 328 persisted Flow v2 opportunities, found {len(ids)}")
-    v1 = get_opportunity_next_actions(ids)
-    v2 = _load_v2_comparison_rows(ids)
-    comparisons = compare_v1_v2_decisions(v1, v2)
-    counts = Counter("agree" if row["action_agrees"] else "different" for row in comparisons)
-    counts["missing_projection"] = sum(not row["projection_present"] for row in comparisons)
-    compare_report = {
-        "generated_at": datetime.now(timezone.utc), "database": identity,
-        "mode_semantics": "V1 returned; persisted V2 observed; no business mutation",
-        "opportunity_count": len(comparisons), "counts": dict(counts), "comparisons": comparisons,
+def run_audit(*, production_readonly_audit: bool) -> dict[str, Any]:
+    # Imports occur only after argparse, so --help cannot initialize DB code.
+    from app.config import settings
+    from app.db import engine
+    from scripts.simulate_blif_flow_v2 import collect
+
+    identity, opportunity_ids, persisted = _database_snapshot(engine)
+    mode = str(settings.blif_flow_v2_mode or "").strip().lower()
+    validate_execution(
+        production_readonly_audit=production_readonly_audit,
+        database=identity["database"], transaction_read_only=identity["transaction_read_only"],
+        mode=mode,
+    )
+    preflight = {**identity, "BLIF_FLOW_V2_MODE": mode,
+                 "production_readonly_opt_in": production_readonly_audit}
+    print(json.dumps(preflight, sort_keys=True), flush=True)
+
+    expected_count = None if production_readonly_audit else 328
+    projection = collect(
+        expected_database=identity["database"], expected_user=identity["user"],
+        require_read_only=production_readonly_audit,
+        expected_opportunity_count=expected_count, require_opportunities=True,
+    )
+    records = {row["opportunity_id"]: row for row in projection["opportunities"]}
+    if set(records) != set(opportunity_ids):
+        raise RuntimeError("factual collector did not return the complete opportunity universe")
+
+    comparisons = []
+    for opportunity_id in opportunity_ids:
+        record = records[opportunity_id]
+        v1, operational = record["v1"], record["safe_v2"]
+        v2 = persisted.get(opportunity_id)
+        classification, reason = classify_semantic_difference(
+            v1_state=v1.get("commercial_stage"), v1_action=v1.get("current_action"),
+            v2_state=(v2 or {}).get("business_state"), v2_action=(v2 or {}).get("business_next_action"),
+            v2_projection_present=v2 is not None,
+            is_duplicate_representation=bool((v2 or {}).get("is_duplicate_representation")),
+            operational_action=operational.get("effective_operational_action"),
+            operational_precedence=operational.get("precedence"),
+            operational_queue=operational.get("effective_operational_queue"),
+            v2_confidence=(v2 or {}).get("confidence"),
+            v2_diagnostic_status=(v2 or {}).get("diagnostic_status"),
+        )
+        comparisons.append({
+            "opportunity_id": opportunity_id, "title": record.get("title"),
+            "customer_name": record.get("customer"), "classification": classification,
+            "classification_reason": reason,
+            "v1": {"state": v1.get("commercial_stage"), "action": v1.get("current_action"),
+                   "queue": v1.get("operational_queue"), "reason_code": v1.get("reason")},
+            "v2": {"business_state": (v2 or {}).get("business_state"),
+                   "business_next_action": (v2 or {}).get("business_next_action"),
+                   "reason_code": (v2 or {}).get("reason_code"),
+                   "diagnostic_status": (v2 or {}).get("diagnostic_status"),
+                   "confidence": (v2 or {}).get("confidence")},
+            "operational_override": {"action": operational.get("effective_operational_action"),
+                                     "queue": operational.get("effective_operational_queue"),
+                                     "precedence": operational.get("precedence")},
+            "material_process_key": (v2 or {}).get("material_process_key"),
+            "canonical_opportunity_id": (v2 or {}).get("canonical_opportunity_id"),
+            "is_duplicate_representation": bool((v2 or {}).get("is_duplicate_representation")),
+            "evidence_refs": (v2 or {}).get("evidence_refs", []),
+        })
+    counts = Counter(row["classification"] for row in comparisons)
+    for name in CLASSIFICATIONS:
+        counts.setdefault(name, 0)
+    real_conflicts = [row for row in comparisons if row["classification"] == "REAL_CONFLICT"]
+    missing = [row for row in comparisons if row["classification"] == "MISSING_PROJECTION"]
+    semantic_report = {
+        "generated_at": datetime.now(timezone.utc), "preflight": preflight,
+        "opportunity_count": len(opportunity_ids), "projection_count": len(persisted),
+        "classification_counts": dict(sorted(counts.items())),
+        "real_conflicts": real_conflicts, "missing_projections": missing,
+        "comparisons": comparisons,
     }
-    _write(COMPARE, compare_report)
+    _write(SEMANTIC, semantic_report)
 
-    inventory = {
-        "generated_at": datetime.now(timezone.utc), "database": identity,
-        "fields": FIELD_INVENTORY,
-        "summary": {
-            "fields_requiring_data_migration": [],
-            "compatibility_only_fields": [row["field"] for row in FIELD_INVENTORY if row["compatibility_required"]],
-            "safe_to_deprecate_after_consumer_migration": [row["field"] for row in FIELD_INVENTORY if row["eventually_deprecatable"]],
-            "historical_followup_timestamps_preserved": historical_timestamps,
-        },
-    }
-    _write(INVENTORY, inventory)
-
-    projection = collect(expected_database=EXPECTED_DATABASE, expected_user=identity["user"], require_read_only=False)
     named = {}
-    for name in ("INSTALBEIRA", "PANORAMIC SUCCESS", "ENGEXICON", "CONSTRURECUP", "X MAT", "RZSOLAR"):
-        matches = [row for row in projection["opportunities"]
-                   if name.casefold() in f"{row.get('title')} {row.get('customer')}".casefold()]
-        named[name] = [{"opportunity_id": row["opportunity_id"], "material_process_key": row["material_process_key"],
-                        "canonical_process_id": row["canonical_process_id"], "business_state": row["safe_v2"]["business_state"],
-                        "effective_action": row["safe_v2"]["effective_operational_action"],
-                        "queue": row["safe_v2"]["effective_operational_queue"],
-                        "duplicate_suppressed": row["safe_v2"]["precedence"] == "duplicate_representation"}
-                       for row in matches]
-    cutover = {
-        "generated_at": datetime.now(timezone.utc), "database": identity,
-        "source_of_truth": [{"concern": concern, "source": source} for concern, source in SOURCE_OF_TRUTH],
-        "data_migration": {"opportunities_requiring_mutation_before_shadow": 0,
-                           "opportunities_requiring_no_mutation_before_shadow": len(ids),
-                           "broad_repair_required": False,
-                           "stage_write_migration_required": False,
-                           "lifecycle_write_migration_required": False,
-                           "followup_timestamp_cleanup_required": False},
-        "mode_contract": {
-            "off": "No Flow v2 derivation or projection writes are triggered by runtime reads.",
-            "shadow": "V1 returned; explicit projection rebuild is additive/idempotent; no tasks, stage, or UI behavior changed.",
-            "compare": "V1 returned; V2 projection read and structured comparison logged; no disagreement writes.",
-            "authoritative": "Disabled and fail-closed. No activation performed.",
+    by_id = {row["opportunity_id"]: row for row in comparisons}
+    for name, opportunity_id in NAMED_IDS.items():
+        row = by_id[opportunity_id]
+        named[name] = row
+    # Use the complete canonical Operations candidate universe, including
+    # preserved standalone obligations, rather than opportunity cards alone.
+    v1_metrics = _operation_totals(projection["v1_totals"])
+    safe_metrics = _operation_totals(projection["safe_v2_totals"])
+    audit = {
+        "generated_at": datetime.now(timezone.utc), "preflight": preflight,
+        "read_only": True, "business_writes": 0, "projection_writes": 0,
+        "opportunity_count": len(opportunity_ids), "projection_count": len(persisted),
+        "canonical_count": sum(not row.get("is_duplicate_representation") for row in persisted.values()),
+        "duplicate_representation_count": sum(bool(row.get("is_duplicate_representation")) for row in persisted.values()),
+        "operations_metrics": {"v1": v1_metrics, "safe_v2": safe_metrics},
+        "semantic_classification_counts": dict(sorted(counts.items())),
+        "authoritative_cutover_blockers": {
+            "real_conflicts": len(real_conflicts), "missing_projections": len(missing),
+            "blocked": bool(real_conflicts or missing),
         },
-        "production_shadow_blockers": [
-            "Production migration 011 presence was not and must not be checked from this environment.",
-            "Deployment must provide an explicit projection rebuild cadence and comparison-log monitoring/retention.",
-        ],
-        "mode_off_deployment_blockers": [],
-        "shadow_compare_activation_prerequisites": [
-            "Install migration 011 through the normal controlled production migration process.",
-            "Run mode=off first, then explicitly rebuild projection in shadow.",
-            "Monitor structured disagreement rates and missing projections before compare enablement.",
-        ],
-        "authoritative_activation_blockers": [
-            "Authoritative switch intentionally raises and has no enabled code path.",
-            "Operations/UI must consume V2 business state/action followed by OperationalEligibility.",
-            "Legacy stage consumers in orders, forecasts, dashboard, workflow guards and opportunity columns require migration or explicit compatibility adapters.",
-            "Explicit operational override precedence must be implemented for CALL_CUSTOMER, due follow-up, SUPPORT, SEND_INFO, manual review and blockers.",
-            "Production shadow/compare observation, rollback criteria and zero-false-negative acceptance must be completed.",
-        ],
-        "legacy_findings_are": "field-level compatibility findings, not required row mutations",
-        "historical_followup_timestamps_preserved": historical_timestamps,
-        "compare_summary": compare_report["counts"], "named_cases": named,
+        "real_conflicts": real_conflicts, "missing_projections": missing,
+        "named_cases": named,
     }
-    _write(CUTOVER, cutover)
-    print(json.dumps({"database": identity, "opportunities": len(ids),
-                      "compare": compare_report["counts"], "historical_timestamps": historical_timestamps,
-                      "mutation_required": 0, "outputs": [str(CUTOVER), str(INVENTORY), str(COMPARE)]}, indent=2))
+    _write(AUDIT, audit)
+    lines = [
+        "BLIF FLOW V2 PRODUCTION SHADOW READ-ONLY AUDIT", "",
+        f"database: {identity['database']}", f"user: {identity['user']}",
+        f"transaction_read_only: {identity['transaction_read_only']}",
+        f"BLIF_FLOW_V2_MODE: {mode}",
+        f"production_readonly_opt_in: {production_readonly_audit}", "",
+        f"opportunities: {len(opportunity_ids)}", f"projections: {len(persisted)}",
+        f"canonical: {audit['canonical_count']}",
+        f"duplicate representations: {audit['duplicate_representation_count']}", "",
+        f"V1 metrics: {json.dumps(v1_metrics, sort_keys=True)}",
+        f"SAFE V2 metrics: {json.dumps(safe_metrics, sort_keys=True)}", "",
+        f"semantic classifications: {json.dumps(dict(sorted(counts.items())), sort_keys=True)}",
+        f"REAL_CONFLICT blockers: {len(real_conflicts)}",
+        f"MISSING_PROJECTION blockers: {len(missing)}",
+        "business writes: 0", "projection writes: 0",
+    ]
+    SUMMARY.write_text("\n".join(lines) + "\n", encoding="utf-8")
+    return audit
+
+
+def main(argv: Sequence[str] | None = None) -> int:
+    # --help exits here before application/database imports or connections.
+    args = build_parser().parse_args(argv)
+    result = run_audit(production_readonly_audit=args.production_readonly_audit)
+    print(json.dumps({
+        "database": result["preflight"]["database"],
+        "opportunities": result["opportunity_count"], "projections": result["projection_count"],
+        "semantic_classifications": result["semantic_classification_counts"],
+        "authoritative_cutover_blockers": result["authoritative_cutover_blockers"],
+        "outputs": [str(AUDIT), str(SEMANTIC), str(SUMMARY)],
+    }, indent=2, sort_keys=True))
+    return 0
 
 
 if __name__ == "__main__":
-    main()
+    raise SystemExit(main())

scripts/rebuild_blif_flow_v2_projection.py

@@ -1,18 +1,84 @@
 #!/usr/bin/env python3
-"""Rebuild additive BLIF Flow v2 projection tables in safe shadow mode."""
+"""Rebuild additive BLIF Flow v2 projection tables with explicit safeguards."""
 from __future__ import annotations
 
+import argparse
 import json
 import os
 import sys
 from pathlib import Path
+from typing import Sequence
+
 
 ROOT = Path(__file__).resolve().parents[1]
 sys.path.insert(0, str(ROOT))
 os.chdir(ROOT)
 
+
+def build_parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument(
+        "--production-shadow", action="store_true",
+        help="explicitly authorize projection-only shadow writes to database clientflow",
+    )
+    return parser
+
+
+def validate_execution(*, production_shadow: bool, mode: str, database: str) -> None:
+    """Validate CLI intent independently of DATABASE_URL inference."""
+    normalized_mode = str(mode or "").strip().lower()
+    if production_shadow:
+        if normalized_mode != "shadow":
+            raise RuntimeError("--production-shadow requires BLIF_FLOW_V2_MODE=shadow")
+        if database != "clientflow":
+            raise RuntimeError(
+                f"--production-shadow requires database 'clientflow', found {database!r}"
+            )
+        return
+    if database == "clientflow":
+        raise RuntimeError("production database requires explicit --production-shadow opt-in")
+
+
+def main(argv: Sequence[str] | None = None) -> int:
+    # argparse handles --help and exits before any application/DB import below.
+    args = build_parser().parse_args(argv)
+
+    from sqlalchemy import text
     from app.blif_flow_v2_projection_service import rebuild_blif_flow_v2_projection
+    from app.config import settings
+    from app.db import engine
+
+    mode = str(settings.blif_flow_v2_mode or "").strip().lower()
+    with engine.connect() as conn:
+        identity = conn.execute(text(
+            "SELECT current_database(), current_user, current_setting('transaction_read_only')"
+        )).one()
+    database, user, transaction_read_only = identity
+    validate_execution(
+        production_shadow=args.production_shadow, mode=mode, database=database,
+    )
+    print(json.dumps({
+        "database": database, "user": user, "blif_flow_v2_mode": mode,
+        "transaction_read_only": transaction_read_only,
+        "production_shadow_opt_in": args.production_shadow,
+    }, sort_keys=True), flush=True)
+
+    if args.production_shadow:
+        result = rebuild_blif_flow_v2_projection(
+            mode="shadow",
+            # Production is deliberately scoped to this invocation; the
+            # module-level default allowlist remains test-only.
+            allowed_databases=frozenset({"clientflow"}),
+            derive_expected_database="clientflow",
+            derive_expected_user=user,
+            expected_opportunity_count=None,
+            require_opportunities=True,
+        )
+    else:
+        result = rebuild_blif_flow_v2_projection()
+    print(json.dumps(result, indent=2, sort_keys=True))
+    return 0
 
 
 if __name__ == "__main__":
-    print(json.dumps(rebuild_blif_flow_v2_projection(), indent=2, sort_keys=True))
+    raise SystemExit(main())

scripts/simulate_blif_flow_v2.py

@@ -99,6 +99,15 @@ def _load(
 ) -> dict[str, Any]:
     with engine.connect() as conn:
         conn = conn.execution_options(isolation_level="AUTOCOMMIT")
+        transaction_started = False
+        try:
+            # A fresh connection commonly reports transaction_read_only=off.
+            # Establish the protected transaction on the same connection used
+            # for every factual read before validating a strict audit.
+            if require_read_only:
+                conn.execute(text("BEGIN READ ONLY"))
+                transaction_started = True
+
             identity = conn.execute(text(
                 "SELECT current_database(), current_user, current_setting('transaction_read_only')"
             )).one()
@@ -106,8 +115,13 @@ def _load(
                 raise RuntimeError(f"refusing unexpected database identity: {identity!r}")
             if require_read_only and identity[2] != "on":
                 raise RuntimeError(f"read-only simulation requires transaction_read_only=on: {identity!r}")
+
+            # Preserve the development/test path's established ordering: check
+            # its identity first, then protect the factual reads themselves.
+            if not require_read_only:
                 conn.execute(text("BEGIN READ ONLY"))
-        try:
+                transaction_started = True
+
             opportunities = [dict(row) for row in conn.execute(text("""
                 SELECT o.*, o.id::text AS id, o.local_customer_id::text,
                        c.name AS linked_customer_name, c.tax_id, c.email AS fiscal_email,
@@ -157,6 +171,7 @@ def _load(
                 WHERE status IN ('open','needs_review','conflict') ORDER BY created_at
             """)).mappings()]
         finally:
+            if transaction_started:
                 conn.execute(text("ROLLBACK"))
     return {
         "identity": {"database": identity[0], "user": identity[1], "transaction_read_only": identity[2]},
@@ -530,6 +545,8 @@ def collect(
     *, expected_database: str = "clientflow_codex_shadow",
     expected_user: str | None = "clientflow_codex",
     require_read_only: bool = True,
+    expected_opportunity_count: int | None = None,
+    require_opportunities: bool = False,
 ) -> dict[str, Any]:
     data = _load(
         expected_database=expected_database,
@@ -537,8 +554,12 @@ def collect(
         require_read_only=require_read_only,
     )
     opportunities = data["opportunities"]
-    if len(opportunities) != 328:
-        raise RuntimeError(f"expected 328 opportunities, found {len(opportunities)}")
+    if expected_opportunity_count is not None and len(opportunities) != expected_opportunity_count:
+        raise RuntimeError(
+            f"expected {expected_opportunity_count} opportunities, found {len(opportunities)}"
+        )
+    if require_opportunities and not opportunities:
+        raise RuntimeError("Flow v2 projection requires at least one opportunity")
     ids = [_s(opp["id"]) for opp in opportunities]
     v1_decisions = get_opportunity_next_actions(ids)
     operations = get_operations_summary(limit=200)

scripts/simulate_blif_flow_v2_authoritative.py

@@ -0,0 +1,226 @@
+#!/usr/bin/env python3
+"""Simulate authoritative BLIF Flow v2 without changing application mode."""
+from __future__ import annotations
+
+import argparse
+import json
+import os
+from pathlib import Path
+import sys
+from typing import Any
+
+JSON_PATH = Path("/tmp/blif_flow_v2_authoritative_simulation.json")
+TEXT_PATH = Path("/tmp/blif_flow_v2_authoritative_operations.txt")
+CURRENT = {"do_now", "review", "blocked", "exception"}
+CLASSIFICATIONS = {
+    "SATISFIED", "SUPERSEDED", "DUPLICATE_SUPPRESSED", "PREMATURE_REMOVED",
+    "REPLACED_BY_CORRECT_ACTION", "LEGACY_ONLY", "UNSAFE_FALSE_NEGATIVE",
+}
+NAMED = {
+    "Instalbeira": ("5c33db95-fab8-477a-bddd-0b9cc8f91302", "CREATE_PROFORMA", "do_now"),
+    "Panoramic": ("fd79b9a1-07e6-4f61-95e8-09eab89c155e", "FOLLOW_UP_CUSTOMER_REVIEW", "do_now"),
+    "ENGEXICON": ("61f1c955-a372-4ea7-b9b0-b8528d74a141", "PREPARE_ORDER", "do_now"),
+    "CONSTRURECUP": ("e3b23ac5-84db-4763-8a31-a684e873032c", "PREPARE_ORDER", "do_now"),
+    "X MAT canonical": ("dc89a466-db24-401b-bfe9-d47644b2d0c8", None, "not_current"),
+    "X MAT duplicate": ("1816a06e-9a69-4a9b-9279-1263156892d3", None, "suppressed"),
+    "RZSOLAR canonical": ("fd221608-e007-4043-a23d-07e0c119a345", "REVIEW_REQUIRED", "review"),
+    "RZSOLAR duplicate": ("434124fb-ac19-4d78-909a-55761d7e8daa", None, "suppressed"),
+}
+
+
+def parse_args(argv: list[str] | None = None) -> argparse.Namespace:
+    parser = argparse.ArgumentParser(
+        description="Read-only V1 versus simulated authoritative BLIF Flow v2 Operations audit."
+    )
+    parser.add_argument(
+        "--production-readonly-simulation", action="store_true",
+        help="Required opt-in when current_database() is clientflow; requires compare mode and a read-only transaction.",
+    )
+    return parser.parse_args(argv)
+
+
+def _load_runtime():
+    """Application imports are intentionally delayed until after argparse."""
+    root = Path(__file__).resolve().parents[1]
+    if str(root) not in sys.path:
+        sys.path.insert(0, str(root))
+    from app.db import engine
+    from app.operations_service import get_operations_summary
+    from app.opportunity_next_action_service import get_opportunity_next_actions_for_mode
+    return engine, get_operations_summary, get_opportunity_next_actions_for_mode
+
+
+def _all(summary: dict[str, Any]) -> list[dict[str, Any]]:
+    if "simulation_all_items" in summary:
+        return list(summary.get("simulation_all_items") or [])
+    return sum((list(summary.get(key) or []) for key in
+                ("work_items", "waiting_items", "backlog_items", "not_current_items")), [])
+
+
+def _metrics(summary: dict[str, Any]) -> dict[str, int]:
+    queues = [str(item.get("operational_queue") or "") for item in _all(summary)]
+    return {"current": sum(q in CURRENT for q in queues), "do_now": queues.count("do_now"),
+            "review": queues.count("review"), "waiting": queues.count("waiting"),
+            "backlog": queues.count("backlog"), "blocked": queues.count("blocked"),
+            "not_current": queues.count("not_current")}
+
+
+def _key(item: dict[str, Any]) -> str:
+    return str(item.get("work_item_key") or item.get("process_key") or item.get("opportunity_id") or item.get("id"))
+
+
+def _classification(old: dict[str, Any], new: dict[str, Any] | None) -> tuple[str, str]:
+    if new is not None:
+        return "REPLACED_BY_CORRECT_ACTION", "Flow v2 selected a different current action for the same work item."
+    decision = old.get("decision") if isinstance(old.get("decision"), dict) else {}
+    code = str(decision.get("reason_code") or old.get("eligibility_reason_code") or "").upper()
+    if "DUPLICATE" in code:
+        return "DUPLICATE_SUPPRESSED", "The local opportunity is a duplicate material representation."
+    if "SATISF" in code or "ANSWERED" in code:
+        return "SATISFIED", "Later factual evidence satisfies the old obligation."
+    if "SUPERSE" in code or "TERMINAL" in code:
+        return "SUPERSEDED", "A later factual state supersedes the legacy card."
+    if "PREMATURE" in code:
+        return "PREMATURE_REMOVED", "The legacy action is premature for the factual state."
+    if old.get("source") in {"task", "opportunity"}:
+        return "LEGACY_ONLY", "The card is supported only by legacy operational representation."
+    return "UNSAFE_FALSE_NEGATIVE", "No safe factual explanation was found for removing this current card."
+
+
+def build_report(
+    *, identity: dict[str, Any], configured_mode: str, v1: dict[str, Any], v2: dict[str, Any],
+    named_decisions: dict[str, dict[str, Any]], missing_projections: int,
+) -> dict[str, Any]:
+    v1_current = {_key(item): item for item in _all(v1) if item.get("operational_queue") in CURRENT}
+    v2_current = {_key(item): item for item in _all(v2) if item.get("operational_queue") in CURRENT}
+    removed, changed = [], []
+    for key, old in v1_current.items():
+        new = v2_current.get(key)
+        if new is not None and new.get("action_code") == old.get("action_code"):
+            continue
+        classification, reason = _classification(old, new)
+        row = {"work_item_key": key, "opportunity_id": old.get("opportunity_id"),
+               "v1_action": old.get("action_code"), "simulated_v2_action": (new or {}).get("action_code"),
+               "classification": classification, "reason": reason,
+               "source_refs": old.get("source_refs") or []}
+        (changed if new is not None else removed).append(row)
+    added = [{"work_item_key": key, "opportunity_id": item.get("opportunity_id"),
+              "action": item.get("action_code"), "source_refs": item.get("source_refs") or []}
+             for key, item in v2_current.items() if key not in v1_current]
+
+    material_counts: dict[str, int] = {}
+    for item in v2_current.values():
+        material = str(item.get("canonical_opportunity_id") or item.get("opportunity_id") or item.get("process_key"))
+        material_counts[material] = material_counts.get(material, 0) + 1
+    duplicates = [{"material_process": key, "count": count} for key, count in material_counts.items() if count > 1]
+
+    named_cases = {}
+    for name, (oid, expected_action, expected_queue) in NAMED.items():
+        decision = named_decisions.get(oid) or {}
+        suppressed = decision.get("suppress_current_card") is True
+        actual_queue = "suppressed" if suppressed else decision.get("operational_queue")
+        actual_action = decision.get("effective_action")
+        passed = actual_action == expected_action and actual_queue == expected_queue
+        named_cases[name] = {"opportunity_id": oid, "expected_action": expected_action,
+                             "expected_queue": expected_queue, "actual_action": actual_action,
+                             "actual_queue": actual_queue, "passed": passed}
+
+    fiscal = []
+    for item in v2_current.values():
+        decision = item.get("decision") if isinstance(item.get("decision"), dict) else {}
+        if item.get("action_code") == "VALIDATE_FISCAL_CUSTOMER":
+            fiscal.append({"opportunity_id": item.get("opportunity_id"),
+                           "business_action": decision.get("business_next_action"),
+                           "reason_code": decision.get("reason_code"),
+                           "reason": decision.get("description")})
+    unsafe = sum(row["classification"] == "UNSAFE_FALSE_NEGATIVE" for row in removed)
+    return {
+        "identity": identity, "configured_mode": configured_mode,
+        "v1_metrics": _metrics(v1), "authoritative_metrics": _metrics(v2),
+        "semantic_changes": len(removed) + len(changed) + len(added),
+        "removed_cards": removed, "added_cards": added, "changed_actions": changed,
+        "duplicate_cards": duplicates, "duplicate_current_cards": len(duplicates),
+        "missing_projections": missing_projections, "unsafe_false_negatives": unsafe,
+        "named_cases": named_cases, "fiscal_validation_cases": fiscal,
+    }
+
+
+def _write_reports(report: dict[str, Any]) -> None:
+    JSON_PATH.write_text(json.dumps(report, indent=2, ensure_ascii=False, default=str) + "\n")
+    lines = ["BLIF Flow v2 authoritative Operations simulation", "",
+             f"Identity: {report['identity']}", f"Configured mode: {report['configured_mode']}",
+             f"V1: {report['v1_metrics']}", f"Simulated authoritative V2: {report['authoritative_metrics']}",
+             f"Removed: {len(report['removed_cards'])}", f"Added: {len(report['added_cards'])}",
+             f"Changed actions: {len(report['changed_actions'])}",
+             f"Duplicate current cards: {report['duplicate_current_cards']}",
+             f"Missing projections: {report['missing_projections']}",
+             f"UNSAFE_FALSE_NEGATIVE: {report['unsafe_false_negatives']}", "", "Named cases:"]
+    lines.extend(f"- {name}: {case}" for name, case in report["named_cases"].items())
+    lines.extend(["", f"Fiscal validation cases: {len(report['fiscal_validation_cases'])}"])
+    TEXT_PATH.write_text("\n".join(lines) + "\n")
+
+
+def exit_code(report: dict[str, Any]) -> int:
+    failed_named = any(not case.get("passed") for case in report.get("named_cases", {}).values())
+    unsafe = int(report.get("unsafe_false_negatives") or 0)
+    missing = int(report.get("missing_projections") or 0)
+    duplicates = int(report.get("duplicate_current_cards") or 0)
+    return 2 if unsafe or missing or duplicates or failed_named else 0
+
+
+def run(args: argparse.Namespace, *, runtime_loader=_load_runtime) -> dict[str, Any]:
+    engine, get_operations, get_decisions = runtime_loader()
+    configured_mode = os.environ.get("BLIF_FLOW_V2_MODE")
+    conn = engine.connect()
+    try:
+        # One explicit transaction and one factual snapshot for identity, V1,
+        # V2, projection completeness, and named-case decisions.
+        conn.exec_driver_sql("BEGIN READ ONLY")
+        identity = dict(conn.exec_driver_sql("""
+            SELECT current_database() AS database, current_user AS user,
+                   current_setting('transaction_read_only') AS transaction_read_only
+        """).mappings().one())
+        production = identity.get("database") == "clientflow"
+        if production and not args.production_readonly_simulation:
+            raise RuntimeError("production simulation requires --production-readonly-simulation")
+        if args.production_readonly_simulation:
+            if identity.get("database") != "clientflow":
+                raise RuntimeError("production simulation requires current_database() = clientflow")
+            if identity.get("transaction_read_only") != "on":
+                raise RuntimeError("production simulation requires transaction_read_only = on")
+            if configured_mode is None:
+                raise RuntimeError("production simulation requires explicit BLIF_FLOW_V2_MODE=compare")
+            if configured_mode.strip().lower() != "compare":
+                raise RuntimeError("production simulation requires BLIF_FLOW_V2_MODE=compare")
+
+        v1 = get_operations(limit=200, flow_mode="off", connection=conn)
+        v2 = get_operations(limit=200, flow_mode="authoritative", connection=conn)
+        projection_counts = conn.exec_driver_sql("""
+            SELECT (SELECT count(*) FROM opportunities) AS opportunities,
+                   (SELECT count(*) FROM opportunity_flow_state_v2) AS projections
+        """).mappings().one()
+        named_ids = [value[0] for value in NAMED.values()]
+        named_decisions = get_decisions(named_ids, flow_mode="authoritative", connection=conn)
+        if production and os.environ.get("BLIF_FLOW_V2_MODE") != configured_mode:
+            raise RuntimeError("configured BLIF_FLOW_V2_MODE changed during simulation")
+        report = build_report(
+            identity=identity, configured_mode=configured_mode or "unset", v1=v1, v2=v2,
+            named_decisions=named_decisions,
+            missing_projections=max(0, int(projection_counts["opportunities"]) - int(projection_counts["projections"])),
+        )
+        _write_reports(report)
+        return report
+    finally:
+        conn.exec_driver_sql("ROLLBACK")
+        conn.close()
+
+
+def main(argv: list[str] | None = None, *, runtime_loader=_load_runtime) -> int:
+    args = parse_args(argv)  # --help exits before runtime_loader/application imports.
+    report = run(args, runtime_loader=runtime_loader)
+    print(json.dumps(report, indent=2, ensure_ascii=False, default=str))
+    return exit_code(report)
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

tests/test_authoritative_operational_adapter.py

@@ -0,0 +1,99 @@
+from datetime import datetime, timedelta, timezone
+
+from app.authoritative_operational_adapter import decide_authoritative_operation
+from app.canonical_operations import canonicalize_operations
+
+
+NOW = datetime(2026, 8, 16, tzinfo=timezone.utc)
+
+
+def projection(state="AWAITING_CUSTOMER", action=None, **extra):
+    return {"opportunity_id": "opp", "canonical_opportunity_id": "opp",
+            "business_state": state, "business_next_action": action,
+            "confidence": "high", **extra}
+
+
+def task(code, due=NOW, **extra):
+    return {"id": code, "status": "pending", "action_code": code, "due_at": due, **extra}
+
+
+def test_business_baselines_and_terminal():
+    for state, action in [("PROFORMA_REQUIRED", "CREATE_PROFORMA"),
+                          ("INVOICE_REQUIRED", "CREATE_INVOICE"),
+                          ("ODOO_ORDER_REQUIRED", "PREPARE_ORDER"),
+                          ("ODOO_ORDER_CREATED", "VALIDATE_ODOO_ORDER")]:
+        got = decide_authoritative_operation(projection(state, action), now=NOW)
+        assert (got.effective_action, got.queue) == (action, "do_now")
+    assert decide_authoritative_operation(projection("COMPLETED"), now=NOW).queue == "not_current"
+
+
+def test_missing_projection_fails_closed_without_v1():
+    got = decide_authoritative_operation(None, now=NOW)
+    assert (got.effective_action, got.queue, got.reason_code) == ("REVIEW_REQUIRED", "review", "MISSING_V2_PROJECTION")
+
+
+def test_due_and_future_customer_followup_semantics():
+    due = decide_authoritative_operation(projection(), obligations=[task("FOLLOW_UP_CUSTOMER_REVIEW")], now=NOW)
+    assert (due.effective_action, due.queue) == ("FOLLOW_UP_CUSTOMER_REVIEW", "do_now")
+    future = decide_authoritative_operation(projection(), obligations=[task("FOLLOW_UP_CUSTOMER_REVIEW", NOW + timedelta(days=1))], now=NOW)
+    assert (future.effective_action, future.queue) == (None, "waiting")
+    assert decide_authoritative_operation(projection(), now=NOW).queue == "waiting"
+
+
+def test_only_active_non_superseded_tasks_override_and_precedence():
+    stale = task("SUPPORT", resolved_at=NOW)
+    assert decide_authoritative_operation(projection("PROFORMA_REQUIRED", "CREATE_PROFORMA"), obligations=[stale], now=NOW).effective_action == "CREATE_PROFORMA"
+    obligations = [task("CALL_CUSTOMER"), task("SEND_INFO"), task("SUPPORT"), task("REVIEW_MANUALLY")]
+    assert decide_authoritative_operation(projection(), obligations=obligations, now=NOW).effective_action == "REVIEW_MANUALLY"
+    assert decide_authoritative_operation(projection(), obligations=[task("SUPPORT")], now=NOW).effective_action == "SUPPORT"
+    assert decide_authoritative_operation(projection(), obligations=[task("SEND_INFO")], now=NOW).effective_action == "SEND_INFO"
+    assert decide_authoritative_operation(projection(), obligations=[task("CALL_CUSTOMER")], now=NOW).effective_action == "CALL_CUSTOMER"
+
+
+def test_fiscal_and_reconciliation_only_block_formal_action():
+    inquiry = decide_authoritative_operation(projection("INQUIRY", "SEND_INFO"), fiscal_complete=False, reconciliation_blocking=True, now=NOW)
+    assert inquiry.effective_action == "SEND_INFO"
+    formal = decide_authoritative_operation(projection("PROFORMA_REQUIRED", "CREATE_PROFORMA"), fiscal_complete=False, now=NOW)
+    assert (formal.effective_action, formal.blocking_action_code) == ("VALIDATE_FISCAL_CUSTOMER", "VALIDATE_FISCAL_CUSTOMER")
+
+
+def test_duplicate_is_suppressed_even_with_pending_legacy_task():
+    decision = decide_authoritative_operation(projection("REVIEW_REQUIRED", "REVIEW_REQUIRED",
+                                                         is_duplicate_representation=True,
+                                                         canonical_opportunity_id="canonical"),
+                                               obligations=[task("REVIEW_MANUALLY")], now=NOW).to_dict()
+    source = {"source": "task", "id": "t", "status": "pending", "action_code": "REVIEW_MANUALLY",
+              "opportunity_id": "opp", "created_at": NOW}
+    assert canonicalize_operations([source], {"opp": decision})["items"] == []
+
+
+def test_authoritative_queue_and_action_survive_canonical_boundary():
+    decision = decide_authoritative_operation(projection(), obligations=[task("FOLLOW_UP_CUSTOMER_REVIEW")], now=NOW).to_dict()
+    source = {"source": "task", "id": "old", "status": "pending", "action_code": "SEND_INVOICE",
+              "opportunity_id": "opp", "created_at": NOW}
+    item = canonicalize_operations([source], {"opp": decision})["items"][0]
+    assert (item["action_code"], item["operational_queue"]) == ("FOLLOW_UP_CUSTOMER_REVIEW", "do_now")
+
+
+def test_named_case_acceptance_decisions():
+    cases = {
+        "5c33db95-fab8-477a-bddd-0b9cc8f91302": ("PROFORMA_REQUIRED", "CREATE_PROFORMA", [], "CREATE_PROFORMA", "do_now"),
+        "fd79b9a1-07e6-4f61-95e8-09eab89c155e": ("AWAITING_CUSTOMER", None, [task("FOLLOW_UP_CUSTOMER_REVIEW")], "FOLLOW_UP_CUSTOMER_REVIEW", "do_now"),
+        "61f1c955-a372-4ea7-b9b0-b8528d74a141": ("ODOO_ORDER_REQUIRED", "PREPARE_ORDER", [], "PREPARE_ORDER", "do_now"),
+        "e3b23ac5-84db-4763-8a31-a684e873032c": ("ODOO_ORDER_REQUIRED", "PREPARE_ORDER", [], "PREPARE_ORDER", "do_now"),
+        "dc89a466-db24-401b-bfe9-d47644b2d0c8": ("COMPLETED", None, [], None, "not_current"),
+        "fd221608-e007-4043-a23d-07e0c119a345": ("REVIEW_REQUIRED", "REVIEW_REQUIRED", [], "REVIEW_REQUIRED", "review"),
+    }
+    for oid, (state, action, obligations, effective, queue) in cases.items():
+        got = decide_authoritative_operation({**projection(state, action), "opportunity_id": oid,
+                                             "canonical_opportunity_id": oid},
+                                            obligations=obligations, now=NOW)
+        assert (got.effective_action, got.queue) == (effective, queue)
+    for duplicate, canonical in [
+        ("1816a06e-9a69-4a9b-9279-1263156892d3", "dc89a466-db24-401b-bfe9-d47644b2d0c8"),
+        ("434124fb-ac19-4d78-909a-55761d7e8daa", "fd221608-e007-4043-a23d-07e0c119a345"),
+    ]:
+        got = decide_authoritative_operation({**projection("REVIEW_REQUIRED", "REVIEW_REQUIRED"),
+                                             "opportunity_id": duplicate, "canonical_opportunity_id": canonical,
+                                             "is_duplicate_representation": True}, now=NOW)
+        assert (got.effective_action, got.queue, got.canonical_opportunity_id) == (None, "not_current", canonical)

tests/test_blif_flow_v2_authoritative_simulation.py

@@ -0,0 +1,138 @@
+from argparse import Namespace
+import importlib.util
+from pathlib import Path
+
+import pytest
+
+
+PATH = Path("scripts/simulate_blif_flow_v2_authoritative.py")
+SPEC = importlib.util.spec_from_file_location("authoritative_simulation", PATH)
+sim = importlib.util.module_from_spec(SPEC)
+SPEC.loader.exec_module(sim)
+
+
+class Result:
+    def __init__(self, row): self.row = row
+    def mappings(self): return self
+    def one(self): return self.row
+
+
+class Connection:
+    def __init__(self, database="clientflow", readonly="on"):
+        self.database, self.readonly, self.sql = database, readonly, []
+    def exec_driver_sql(self, sql):
+        self.sql.append(sql.strip())
+        if "current_database" in sql:
+            return Result({"database": self.database, "user": "audit", "transaction_read_only": self.readonly})
+        if "count(*) FROM opportunities" in sql:
+            return Result({"opportunities": 328, "projections": 328})
+        return Result({})
+    def close(self): pass
+
+
+class Engine:
+    def __init__(self, connection): self.connection = connection
+    def connect(self): return self.connection
+
+
+def named_decisions(fail=False):
+    result = {}
+    for name, (oid, action, queue) in sim.NAMED.items():
+        result[oid] = {"effective_action": action,
+                       "operational_queue": "not_current" if queue == "suppressed" else queue,
+                       "suppress_current_card": queue == "suppressed"}
+    if fail:
+        result[sim.NAMED["Instalbeira"][0]]["effective_action"] = "SEND_INVOICE"
+    return result
+
+
+def runtime(connection, *, fail_named=False, modes=None):
+    empty = {"work_items": [], "waiting_items": [], "backlog_items": [], "not_current_items": []}
+    def operations(**kwargs):
+        assert kwargs["connection"] is connection
+        if modes is not None:
+            modes.append(kwargs["flow_mode"])
+        return empty
+    def decisions(ids, **kwargs):
+        assert kwargs == {"flow_mode": "authoritative", "connection": connection}
+        return named_decisions(fail_named)
+    return lambda: (Engine(connection), operations, decisions)
+
+
+def test_help_performs_zero_db_work():
+    called = False
+    def loader():
+        nonlocal called
+        called = True
+        raise AssertionError("DB/application runtime must not load")
+    with pytest.raises(SystemExit) as exc:
+        sim.main(["--help"], runtime_loader=loader)
+    assert exc.value.code == 0 and called is False
+
+
+def test_production_requires_explicit_flag(monkeypatch):
+    monkeypatch.setenv("BLIF_FLOW_V2_MODE", "compare")
+    with pytest.raises(RuntimeError, match="requires --production"):
+        sim.run(Namespace(production_readonly_simulation=False), runtime_loader=runtime(Connection()))
+
+
+@pytest.mark.parametrize("database,readonly,mode,message", [
+    ("test_db", "on", "compare", "current_database"),
+    ("clientflow", "off", "compare", "transaction_read_only"),
+    ("clientflow", "on", None, "explicit BLIF_FLOW"),
+    ("clientflow", "on", "off", "BLIF_FLOW_V2_MODE=compare"),
+    ("clientflow", "on", "shadow", "BLIF_FLOW_V2_MODE=compare"),
+    ("clientflow", "on", "authoritative", "BLIF_FLOW_V2_MODE=compare"),
+])
+def test_production_guards(monkeypatch, database, readonly, mode, message):
+    if mode is None:
+        monkeypatch.delenv("BLIF_FLOW_V2_MODE", raising=False)
+    else:
+        monkeypatch.setenv("BLIF_FLOW_V2_MODE", mode)
+    with pytest.raises(RuntimeError, match=message):
+        sim.run(Namespace(production_readonly_simulation=True),
+                runtime_loader=runtime(Connection(database, readonly)))
+
+
+def test_safe_simulation_uses_explicit_modes_without_changing_config_and_has_no_db_writes(monkeypatch, tmp_path):
+    monkeypatch.setenv("BLIF_FLOW_V2_MODE", "compare")
+    monkeypatch.setattr(sim, "JSON_PATH", tmp_path / "report.json")
+    monkeypatch.setattr(sim, "TEXT_PATH", tmp_path / "report.txt")
+    connection = Connection()
+    modes = []
+    report = sim.run(Namespace(production_readonly_simulation=True), runtime_loader=runtime(connection, modes=modes))
+    assert report["configured_mode"] == "compare"
+    assert sim.exit_code(report) == 0
+    assert [sql for sql in connection.sql if sql.split(None, 1)[0].upper() in {"INSERT", "UPDATE", "DELETE", "MERGE"}] == []
+    assert connection.sql[0] == "BEGIN READ ONLY" and connection.sql[-1] == "ROLLBACK"
+    assert modes == ["off", "authoritative"]
+
+
+def safe_report():
+    return sim.build_report(identity={}, configured_mode="compare",
+                            v1={"work_items": []}, v2={"work_items": []},
+                            named_decisions=named_decisions(), missing_projections=0)
+
+
+def test_unsafe_false_negative_is_nonzero():
+    report = safe_report()
+    report["unsafe_false_negatives"] = 1
+    assert sim.exit_code(report) != 0
+
+
+def test_missing_projection_is_nonzero():
+    report = safe_report()
+    report["missing_projections"] = 1
+    assert sim.exit_code(report) != 0
+
+
+def test_duplicate_current_card_is_nonzero():
+    report = safe_report()
+    report["duplicate_current_cards"] = 1
+    assert sim.exit_code(report) != 0
+
+
+def test_named_case_failure_is_nonzero():
+    report = sim.build_report(identity={}, configured_mode="compare", v1={"work_items": []},
+                              v2={"work_items": []}, named_decisions=named_decisions(True), missing_projections=0)
+    assert sim.exit_code(report) != 0

tests/test_blif_flow_v2_cutover.py

@@ -76,10 +76,9 @@ def test_duplicate_representation_stays_suppressed():
     assert suppressed.effective_operational_queue == "not_current"
 
 
-def test_authoritative_mode_is_fail_closed(monkeypatch):
+def test_authoritative_empty_bulk_is_safe(monkeypatch):
     monkeypatch.setattr(service.settings, "blif_flow_v2_mode", "authoritative")
-    with pytest.raises(RuntimeError, match="disabled"):
-        service.get_opportunity_next_actions([])
+    assert service.get_opportunity_next_actions([]) == {}
 
 
 def test_shadow_mode_has_no_decision_side_effect(monkeypatch):

tests/test_blif_flow_v2_production_readonly_audit.py

@@ -0,0 +1,110 @@
+from inspect import getsource
+
+import pytest
+
+import scripts.audit_blif_flow_v2_cutover as audit
+
+
+def classify(**overrides):
+    values = dict(
+        v1_state="INFO_SENT", v1_action="SEND_INFO",
+        v2_state="INQUIRY", v2_action="SEND_INFO",
+        v2_projection_present=True, is_duplicate_representation=False,
+        operational_action="SEND_INFO", operational_precedence="business_transition",
+        operational_queue="do_now", v2_confidence="high", v2_diagnostic_status="clear",
+    )
+    values.update(overrides)
+    return audit.classify_semantic_difference(**values)[0]
+
+
+def test_help_performs_zero_database_work(monkeypatch, capsys):
+    monkeypatch.setattr(audit, "run_audit", lambda **kwargs: pytest.fail("help reached DB audit"))
+    with pytest.raises(SystemExit) as exc:
+        audit.main(["--help"])
+    assert exc.value.code == 0
+    assert "--production-readonly-audit" in capsys.readouterr().out
+
+
+def test_production_requires_explicit_opt_in():
+    with pytest.raises(RuntimeError, match="explicit --production-readonly-audit"):
+        audit.validate_execution(production_readonly_audit=False, database="clientflow",
+                                 transaction_read_only="on", mode="shadow")
+
+
+def test_production_database_must_be_clientflow():
+    with pytest.raises(RuntimeError, match="requires database 'clientflow'"):
+        audit.validate_execution(production_readonly_audit=True, database="clientflow_codex_test",
+                                 transaction_read_only="on", mode="shadow")
+
+
+def test_production_transaction_must_be_read_only():
+    with pytest.raises(RuntimeError, match="transaction_read_only=on"):
+        audit.validate_execution(production_readonly_audit=True, database="clientflow",
+                                 transaction_read_only="off", mode="shadow")
+
+
+def test_production_audit_contains_no_sql_writes():
+    source = (getsource(audit._database_snapshot) + getsource(audit.run_audit)).upper()
+    for verb in ("INSERT ", "UPDATE ", "DELETE ", "CREATE ", "ALTER ", "DROP ", "TRUNCATE "):
+        assert verb not in source
+    assert 'BEGIN READ ONLY' in source
+
+
+@pytest.mark.parametrize("mode", ["shadow", "compare"])
+def test_shadow_and_compare_are_allowed(mode):
+    audit.validate_execution(production_readonly_audit=True, database="clientflow",
+                             transaction_read_only="on", mode=mode)
+
+
+@pytest.mark.parametrize("mode", ["", "off", "authoritative"])
+def test_unsafe_production_modes_are_refused(mode):
+    with pytest.raises(RuntimeError, match="requires BLIF_FLOW_V2_MODE=shadow or compare"):
+        audit.validate_execution(production_readonly_audit=True, database="clientflow",
+                                 transaction_read_only="on", mode=mode)
+
+
+def test_missing_projection_is_reported():
+    assert classify(v2_projection_present=False, v2_state=None, v2_action=None) == "MISSING_PROJECTION"
+
+
+def test_semantically_equivalent_is_classified():
+    assert classify() == "SEMANTICALLY_EQUIVALENT"
+    assert classify(v1_action="NO_ACTION", v2_state="COMPLETED", v2_action=None,
+                    operational_action=None, operational_queue="not_current") == "SEMANTICALLY_EQUIVALENT"
+
+
+def test_operational_override_is_classified():
+    assert classify(v1_action="FOLLOW_UP_CUSTOMER_REVIEW", v2_state="AWAITING_CUSTOMER",
+                    v2_action=None, operational_action="FOLLOW_UP_CUSTOMER_REVIEW",
+                    operational_precedence="due_followup") == "V1_OPERATIONAL_OVERRIDE"
+
+
+def test_v2_correction_is_classified():
+    assert classify(v1_action="SEND_INVOICE", v2_state="PROFORMA_REQUIRED",
+                    v2_action="CREATE_PROFORMA") == "V2_CORRECTS_V1"
+    assert classify(v1_action="REVIEW_RECONSTRUCTED_PROCESS", v2_state="REVIEW_REQUIRED",
+                    v2_action="REVIEW_REQUIRED", is_duplicate_representation=True) == "V2_CORRECTS_V1"
+
+
+def test_legacy_only_is_classified():
+    assert classify(v1_action="CREATE_JASMIN_QUOTE", v2_state="AWAITING_CUSTOMER",
+                    v2_action=None, v2_confidence="medium") == "LEGACY_ONLY"
+
+
+def test_real_conflict_is_classified():
+    assert classify(v1_action="UNMAPPED_ACTION", v2_state="INQUIRY",
+                    v2_action=None, v2_confidence="medium",
+                    v2_diagnostic_status="ambiguous", operational_action=None) == "REAL_CONFLICT"
+
+
+def test_named_case_ids_and_expected_semantics_are_frozen():
+    assert audit.NAMED_IDS == {
+        "INSTALBEIRA": "5c33db95-fab8-477a-bddd-0b9cc8f91302",
+        "PANORAMIC": "fd79b9a1-07e6-4f61-95e8-09eab89c155e",
+        "ENGEXICON": "61f1c955-a372-4ea7-b9b0-b8528d74a141",
+        "CONSTRURECUP": "e3b23ac5-84db-4763-8a31-a684e873032c",
+        "X_MAT_CANONICAL": "dc89a466-db24-401b-bfe9-d47644b2d0c8",
+        "X_MAT_DUPLICATE": "1816a06e-9a69-4a9b-9279-1263156892d3",
+        "RZSOLAR_CANONICAL": "fd221608-e007-4043-a23d-07e0c119a345",
+        "RZSOLAR_DUPLICATE": "434124fb-ac19-4d78-909a-55761d7e8daa",
+    }

tests/test_blif_flow_v2_production_shadow_rebuild.py

@@ -0,0 +1,104 @@
+from inspect import getsource
+
+import pytest
+
+import app.blif_flow_v2_projection_service as projection
+import scripts.rebuild_blif_flow_v2_projection as cli
+import scripts.simulate_blif_flow_v2 as simulator
+
+
+def test_help_performs_no_rebuild(monkeypatch, capsys):
+    monkeypatch.setattr(cli, "validate_execution", lambda **kwargs: pytest.fail("help reached execution"))
+    with pytest.raises(SystemExit) as exc:
+        cli.main(["--help"])
+    assert exc.value.code == 0
+    assert "--production-shadow" in capsys.readouterr().out
+
+
+def test_default_execution_does_not_authorize_production():
+    with pytest.raises(RuntimeError, match="explicit --production-shadow"):
+        cli.validate_execution(production_shadow=False, mode="shadow", database="clientflow")
+
+
+@pytest.mark.parametrize("mode", ["", "off", "compare", "authoritative"])
+def test_production_shadow_requires_exact_shadow_mode(mode):
+    with pytest.raises(RuntimeError, match="requires BLIF_FLOW_V2_MODE=shadow"):
+        cli.validate_execution(production_shadow=True, mode=mode, database="clientflow")
+
+
+@pytest.mark.parametrize("database", ["clientflow_codex_test", "clientflow_codex_shadow", "other"])
+def test_production_shadow_requires_exact_production_database(database):
+    with pytest.raises(RuntimeError, match="requires database 'clientflow'"):
+        cli.validate_execution(production_shadow=True, mode="shadow", database=database)
+
+
+def test_production_shadow_explicit_proof_is_accepted():
+    cli.validate_execution(production_shadow=True, mode="shadow", database="clientflow")
+
+
+def test_global_write_allowlist_remains_test_only():
+    assert projection.WRITE_DATABASE_ALLOWLIST == frozenset({"clientflow_codex_test"})
+    assert "clientflow" not in projection.WRITE_DATABASE_ALLOWLIST
+
+
+def test_factual_read_transaction_remains_read_only():
+    source = getsource(simulator._load)
+    assert 'conn.execute(text("BEGIN READ ONLY"))' in source
+    assert 'conn.execute(text("ROLLBACK"))' in source
+
+
+def test_projection_writer_only_mutates_two_additive_tables():
+    source = getsource(projection.rebuild_blif_flow_v2_projection).upper()
+    assert projection.PROJECTION_WRITE_TABLES == {
+        "opportunity_flow_state_v2", "opportunity_flow_transitions"}
+    assert "INSERT INTO OPPORTUNITY_FLOW_TRANSITIONS" in source
+    assert "INSERT INTO OPPORTUNITY_FLOW_STATE_V2" in source
+    for forbidden in ("OPPORTUNITIES", "TASKS", "MESSAGES", "COMMUNICATIONS",
+                      "COMMERCIAL_DOCUMENTS", "CUSTOMERS", "PAYMENTS", "OPERATION_LINKS"):
+        assert f"INSERT INTO {forbidden}" not in source
+        assert f"UPDATE {forbidden}" not in source
+        assert f"DELETE FROM {forbidden}" not in source
+
+
+def test_production_derivation_has_no_fixed_328_requirement(monkeypatch):
+    captured = {}
+
+    def fake_collect(**kwargs):
+        captured.update(kwargs)
+        return {"opportunities": [{"opportunity_id": "one"}]}
+
+    monkeypatch.setattr(simulator, "collect", fake_collect)
+    rows = projection._derive_all(
+        expected_database="clientflow", expected_user="runtime-role",
+        expected_opportunity_count=None, require_opportunities=True,
+    )
+    assert rows == [{"opportunity_id": "one"}]
+    assert captured["expected_opportunity_count"] is None
+    assert captured["require_opportunities"] is True
+
+
+def test_snapshot_expected_count_validation_is_still_available(monkeypatch):
+    monkeypatch.setattr(simulator, "_load", lambda **kwargs: {"opportunities": [object()]})
+    with pytest.raises(RuntimeError, match="expected 328 opportunities, found 1"):
+        simulator.collect(expected_opportunity_count=328)
+
+
+def test_empty_production_universe_is_rejected(monkeypatch):
+    monkeypatch.setattr(simulator, "collect", lambda **kwargs: {"opportunities": []})
+    assert projection._derive_all(
+        expected_database="clientflow", expected_user=None,
+        expected_opportunity_count=None, require_opportunities=True,
+    ) == []
+    with pytest.raises(RuntimeError, match="at least one opportunity"):
+        projection.rebuild_blif_flow_v2_projection(
+            mode="shadow", derived_rows=[], require_opportunities=True,
+        )
+
+
+def test_existing_test_defaults_remain_guarded(monkeypatch):
+    captured = {}
+    monkeypatch.setattr(simulator, "collect", lambda **kwargs: captured.update(kwargs) or {"opportunities": []})
+    projection._derive_all()
+    assert captured["expected_database"] == "clientflow_codex_test"
+    assert captured["expected_user"] == "clientflow_codex_test"
+    assert captured["expected_opportunity_count"] == 328

tests/test_blif_flow_v2_readonly_transaction_order.py

@@ -0,0 +1,125 @@
+from contextlib import nullcontext
+
+import pytest
+
+import scripts.simulate_blif_flow_v2 as simulator
+
+
+class _Result:
+    def __init__(self, *, identity=None, rows=()):
+        self._identity = identity
+        self._rows = rows
+
+    def one(self):
+        return self._identity
+
+    def mappings(self):
+        return self._rows
+
+
+class _Connection:
+    def __init__(self, database="clientflow", user="clientflow"):
+        self.database = database
+        self.user = user
+        self.read_only = False
+        self.statements = []
+        self.identity_observations = []
+
+    def execution_options(self, **_kwargs):
+        return self
+
+    def execute(self, statement):
+        sql = " ".join(str(statement).split())
+        self.statements.append(sql)
+        upper = sql.upper()
+        if upper == "BEGIN READ ONLY":
+            self.read_only = True
+            return _Result()
+        if upper == "ROLLBACK":
+            self.read_only = False
+            return _Result()
+        if upper.startswith(("INSERT ", "UPDATE ", "DELETE ")):
+            if self.read_only:
+                raise RuntimeError("cannot execute write in a read-only transaction")
+            return _Result()
+        if "CURRENT_DATABASE()" in upper:
+            identity = (self.database, self.user, "on" if self.read_only else "off")
+            self.identity_observations.append(identity)
+            return _Result(identity=identity)
+        return _Result(rows=())
+
+
+class _Engine:
+    def __init__(self, connection):
+        self.connection = connection
+
+    def connect(self):
+        return nullcontext(self.connection)
+
+
+def _load(monkeypatch, connection, **kwargs):
+    monkeypatch.setattr(simulator, "engine", _Engine(connection))
+    return simulator._load(**kwargs)
+
+
+def test_strict_load_establishes_read_only_before_same_connection_identity_and_reads(monkeypatch):
+    connection = _Connection()
+    assert connection.read_only is False  # normal fresh-connection state
+
+    result = _load(
+        monkeypatch, connection, expected_database="clientflow",
+        expected_user="clientflow", require_read_only=True,
+    )
+
+    assert connection.statements[0] == "BEGIN READ ONLY"
+    assert "CURRENT_DATABASE()" in connection.statements[1].upper()
+    assert connection.identity_observations == [("clientflow", "clientflow", "on")]
+    assert result["identity"]["transaction_read_only"] == "on"
+    assert connection.statements[2].upper().startswith("SELECT O.*")
+    assert connection.statements[-1] == "ROLLBACK"
+
+
+@pytest.mark.parametrize(
+    ("database", "user", "expected_database", "expected_user"),
+    [
+        ("wrong", "clientflow", "clientflow", "clientflow"),
+        ("clientflow", "wrong", "clientflow", "clientflow"),
+    ],
+)
+def test_strict_load_refuses_wrong_identity_before_factual_reads(
+    monkeypatch, database, user, expected_database, expected_user,
+):
+    connection = _Connection(database=database, user=user)
+    with pytest.raises(RuntimeError, match="unexpected database identity"):
+        _load(
+            monkeypatch, connection, expected_database=expected_database,
+            expected_user=expected_user, require_read_only=True,
+        )
+
+    assert connection.statements[0] == "BEGIN READ ONLY"
+    assert len([sql for sql in connection.statements if sql.upper().startswith("SELECT")]) == 1
+    assert connection.statements[-1] == "ROLLBACK"
+
+
+def test_strict_factual_transaction_rejects_writes():
+    connection = _Connection()
+    connection.execute("BEGIN READ ONLY")
+    with pytest.raises(RuntimeError, match="read-only transaction"):
+        connection.execute("UPDATE opportunities SET stage = 'forbidden'")
+    assert connection.read_only is True
+
+
+def test_non_strict_load_preserves_identity_then_read_only_collection_order(monkeypatch):
+    connection = _Connection(database="clientflow_codex_test", user="clientflow_codex_test")
+    result = _load(
+        monkeypatch, connection, expected_database="clientflow_codex_test",
+        expected_user="clientflow_codex_test", require_read_only=False,
+    )
+
+    assert "CURRENT_DATABASE()" in connection.statements[0].upper()
+    assert connection.identity_observations == [
+        ("clientflow_codex_test", "clientflow_codex_test", "off")
+    ]
+    assert connection.statements[1] == "BEGIN READ ONLY"
+    assert result["identity"]["transaction_read_only"] == "off"
+    assert connection.statements[-1] == "ROLLBACK"

tests/test_opportunity_next_action_bulk.py

@@ -8,6 +8,12 @@ import app.opportunity_next_action_service as service
 from app.domain.opportunity_flow import build_opportunity_evidence
 
 
+@pytest.fixture(autouse=True)
+def _v1_contract_mode(monkeypatch):
+    """These tests measure the legacy bulk contract, independently of env mode."""
+    monkeypatch.setattr(service.settings, "blif_flow_v2_mode", "off")
+
+
 class _Result:
     def __init__(self, rows):
         self._rows = rows