feat(p21): the PROJECT carries the price, so a run against a road normal can be anchored

Four paid stress rounds ran entirely UN-ANCHORED, all of them, because the one file
loader reads cost-baseline.json out of the BUNDLE and no vegnormal ships one: N100,
N200, N500 and R761 are knowledge, and knowledge carries requirements, never amounts.
The validator's stage 0 -- the one stage that tells an invented cost line from a line
this project actually buys -- was skipped in every single run, so "validated" could not
mean what it says. P20 G1/G2 measured real R761 process numbers (12.11 three times on
Soraasen, 1.1.1 on Lindaas) validating with amounts nobody had anywhere.

--cost-baseline FILE is PM decision (e), taken over the three alternatives P20 wrote
down. A LOADED object, never a path (prepass_payload's rule): the CLI owns the file and
loads it ONCE, so the notice, the stamp and every base of an --across-bundle pass all
descend from one read. ONE parse, two doors -- load_cost_baseline delegates to
load_cost_baseline_file -- while safe_resolve stays on the bundle door alone, because a
project's own schedule is legitimately outside every base. No tolerant twin: this path
exists only because an operator NAMED a file.

DEL B: five anchored context sets, a1-a3 with their line and a4 with none, so stage 0 is
what catches the falsification arm. THE ORDER'S OWN ARM (h) WAS FELLED BY MEASUREMENT:
"no baseline code is a requirement number the base declares" is measured 0 of 4 on the
project-coded sets and 5 of 5 on kontrakt-sorasen -- which is what R761 Prosesskoden IS,
a bill of quantities priced BY process code. The complement keeps both, and the order's
own mutation still bites.

DEL B3: the judge reports anchored (off the run's own stamp), priced per row, and WHICH
falsifier caught the falsification arm.

Load-bearing MEASURED, five mutations all red against the WHOLE suite, green control
1850/5 (from 1809/5, superset, 0 removed), golden byte-unchanged:
A3(i) the flag is read but the baseline is unused (3 red) . A3(ii) only the first base
gets it (1) . A3(iii) report_forbidden drops it (1) . B2(i) a4 gets a line (1, arm (g)
alone) . B2(ii) a code swapped to 12.11 (2, arms (f) and (h)).

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

tests/test_prose_code_form_loadbearing.py

@@ -209,3 +209,64 @@ def test_the_classification_is_reported_and_is_the_gates_own() -> None:
         verdict = validate_proposal(_proposal(code), grounding=Grounding(_OFFERING + (code,)))
         refused = isinstance(verdict, Rejection) and "has no identifier form" in verdict.reason
         assert refused == (kind == "prose"), (code, kind, verdict)
+
+
+# --------------------------------------------------------------------------------------------
+# P21 B3: ``rejection_stage`` — which falsifier wrote a reason. A REPORT, never a gate: nothing
+# branches on it, so an unrecognised sentence costs a label rather than a verdict. It lives beside
+# the sentences it keys on, so the classifier and the wordings cannot drift apart (kø-(p)).
+# --------------------------------------------------------------------------------------------
+
+
+def test_p21_rejection_stage_names_each_stage_from_its_own_sentence() -> None:
+    from portfolio_optimiser.validator import rejection_stage
+
+    assert (
+        rejection_stage("unknown cost code 'X': not in project p's cost baseline (3 known codes)")
+        == "stage0-baseline"
+    )
+    assert (
+        rejection_stage(
+            "quantity 5 for cost code 'X' is outside the 5.0% tolerance around the baseline "
+            "quantity 9"
+        )
+        == "stage0-baseline"
+    )
+    assert (
+        rejection_stage("ungrounded identifier 'X': it appears nowhere in the input (9 chars)")
+        == "stage0b-grounding"
+    )
+    assert rejection_stage("claimed saving 9 exceeds P90 feasible 4") == "stage4-p90"
+    assert (
+        rejection_stage("claimed saving 9 exceeds the nominal feasible 4 at the items' stated")
+        == "stage4b-nominal"
+    )
+    assert (
+        rejection_stage("claimed 9 exceeds the energy_efficiency method cap 4 (stricter)")
+        == "stage5-method-cap"
+    )
+    # The honest answer for a sentence this module did not write — a label, never a verdict.
+    assert rejection_stage("something else entirely") == "other"
+
+
+def test_p21_rejection_stage_is_keyed_on_the_sentences_the_validator_emits() -> None:
+    """The control: the markers are not a private paraphrase but the text stage 0 really writes.
+
+    Without it the classifier could key on wording nothing emits and every arm above would still
+    be green — the vacuous-gate class, on a reporter.
+    """
+    from portfolio_optimiser.ir import AffectedItem, CostBaseline, CostBaselineLine, SavingsProposal
+    from portfolio_optimiser.validator import Rejection, rejection_stage, validate_proposal
+
+    baseline = CostBaseline(
+        project_id="proj", items={"REAL-1": CostBaselineLine(quantity=10.0, unit_cost=100.0)}
+    )
+    proposal = SavingsProposal(
+        project_id="proj",
+        measure="m",
+        affected_items=[AffectedItem(code="FAKE-1", quantity=10.0, unit_cost=100.0)],
+        claimed_saving_nok=100.0,
+    )
+    outcome = validate_proposal(proposal, baseline=baseline)
+    assert isinstance(outcome, Rejection)
+    assert rejection_stage(outcome.reason) == "stage0-baseline"