Address PR #356 CI review round 14 (1 P1 guide)

dCDH "spillover designs" claim removed. The earlier §4.3 bullet said
ChaisemartinDHaultfoeuille was "robust to non-absorbing treatment and
to spillover designs," but REGISTRY.md §ChaisemartinDHaultfoeuille
only documents support for non-absorbing / reversible treatment — the
estimator assumes SUTVA like every other DiD estimator in the suite.
The same guide's §7 glossary already defines SUTVA as ruling out
spillovers, so the §4.3 wording was internally inconsistent and would
have misrouted interference designs to dCDH.

Rewrote the §4.3 bullet to state the actually-supported contract
(non-absorbing / reversible treatment via DID_M / DID_l) and added
an explicit note that interference / between-unit spillovers are not
supported natively.

Regression test: `tests/test_profile_panel.py` extended with a
forbidden-phrase check that fails if the autonomous guide re-advertises
dCDH as "robust to spillover", "interference-robust", "supports
spillover", or "and to spillover".

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: igerber

diff_diff/guides/llms-autonomous.txt

@@ -339,7 +339,9 @@ estimators:
   coefficients. Requires a never-treated cohort (`fit` raises a
   `ValueError` when none exists).
 - `ChaisemartinDHaultfoeuille` - DID_M / DID_l estimators robust to
-  non-absorbing treatment (see §4.5) and to spillover designs.
+  non-absorbing / reversible treatment (see §4.5). Interference /
+  between-unit spillovers are not supported natively - SUTVA is
+  assumed like every other DiD estimator in the suite.
 - `ImputationDiD` (Borusyak, Jaravel, Spiess) - imputation-based,
   efficient under homoskedasticity, produces an imputation-based
   residual at the observation level.

tests/test_profile_panel.py

@@ -729,6 +729,22 @@ def test_guide_api_strings_resolve_against_public_api():
         "intolerant estimators"
     )
 
+    # ChaisemartinDHaultfoeuille handles non-absorbing / reversible
+    # treatment; SUTVA is still assumed (no native interference or
+    # spillover support per REGISTRY.md). Guard against the guide
+    # drifting back to advertising dCDH as "robust to spillover
+    # designs" or similar.
+    for phrase in (
+        "robust to spillover",
+        "interference-robust",
+        "supports spillover",
+        "and to spillover",
+    ):
+        assert phrase not in text, (
+            f"Guide must not advertise unsupported dCDH capability "
+            f"{phrase!r}: SUTVA is assumed across the estimator suite."
+        )
+
 
 def test_min_pre_post_use_per_unit_observed_support():
     """On an unbalanced panel where one treated unit is missing its