Address PR #372 R4 review (1 P2, 1 P3) + rebase onto main

(Branch rebased onto origin/main first - 32 landed commits replayed cleanly,
zero file overlap with the docs-refresh changeset.)

P2 - llms-practitioner.txt Step 6 inconsistent on dCDH HonestDiD support
- The high-level workflow surfaces (README, bundled llms.txt) and the
  practitioner guide's estimator tree (L177-180) all advertise
  compute_honest_did(results) on MultiPeriodDiD / CallawaySantAnna /
  ChaisemartinDHaultfoeuille, but Step 6 still said "MultiPeriodDiD and
  CallawaySantAnna only".
- Update Step 6 to add dCDH with the registry caveats: requires L_max >= 1,
  bounds use placebo estimates as pre-period coefficients (not standard
  event-study pre-treatment coefficients), uses diagonal variance (no full
  VCV available), emits UserWarning. Cross-references the REGISTRY.md
  "Note (HonestDiD integration)" entry under ChaisemartinDHaultfoeuille
  for the full caveat list.

P3 - docs/doc-deps.yaml unsectioned llms.txt entries
- Add `section: "Estimators"` to the 5 remaining unsectioned
  `diff_diff/guides/llms.txt` mappings (estimators.py, staggered.py,
  chaisemartin_dhaultfoeuille.py, results.py, __init__.py). Now every
  llms.txt mapping in doc-deps.yaml carries a section label, matching the
  new "first-class documentation surface" contract.

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

Author: igerber

diff_diff/guides/llms-practitioner.txt

@@ -293,11 +293,19 @@ print(results.summary())
 
 This step is CRITICAL and most often skipped. Run at least one of:
 
-### HonestDiD (Rambachan & Roth 2023) — recommended
+### HonestDiD (Rambachan & Roth 2023) - recommended
 Bounds on the treatment effect under violations of parallel trends.
-Works with MultiPeriodDiD and CallawaySantAnna results only. For CS,
-requires `aggregate='event_study'` or `aggregate='all'` so that event
-study effects are available.
+Works with MultiPeriodDiD, CallawaySantAnna, and ChaisemartinDHaultfoeuille
+(dCDH) results.
+
+- For CS: requires `aggregate='event_study'` or `aggregate='all'` so that
+  event study effects are available.
+- For dCDH: requires `L_max >= 1` (multi-horizon mode). Bounds use placebo
+  estimates `DID^{pl}_l` as pre-period coefficients rather than standard
+  event-study pre-treatment coefficients, and use diagonal variance (no
+  full VCV available for dCDH); a `UserWarning` is emitted at runtime.
+  See `docs/methodology/REGISTRY.md` "Note (HonestDiD integration)" under
+  ChaisemartinDHaultfoeuille for the full caveats.
 
 ```python
 from diff_diff import compute_honest_did

docs/doc-deps.yaml

@@ -102,6 +102,7 @@ sources:
         section: "DifferenceInDifferences"
         type: user_guide
       - path: diff_diff/guides/llms.txt
+        section: "Estimators"
         type: user_guide
       - path: docs/choosing_estimator.rst
         type: user_guide
@@ -145,6 +146,7 @@ sources:
         section: "CallawaySantAnna"
         type: user_guide
       - path: diff_diff/guides/llms.txt
+        section: "Estimators"
         type: user_guide
       - path: docs/choosing_estimator.rst
         type: user_guide
@@ -314,6 +316,7 @@ sources:
         section: "ChaisemartinDHaultfoeuille"
         type: user_guide
       - path: diff_diff/guides/llms.txt
+        section: "Estimators"
         type: user_guide
       - path: docs/choosing_estimator.rst
         type: user_guide
@@ -690,6 +693,7 @@ sources:
         section: "Results API"
         type: user_guide
       - path: diff_diff/guides/llms.txt
+        section: "Estimators"
         type: user_guide
       - path: docs/methodology/REGISTRY.md
         section: "SyntheticDiD"
@@ -758,6 +762,7 @@ sources:
     drift_risk: low
     docs:
       - path: diff_diff/guides/llms.txt
+        section: "Estimators"
         type: user_guide
         note: "Public API surface"