docs(robustness): first backfill batch — 3 [@bibkey] annotations in sensitivity_frontier

First real exercise of tools/suggest_bibkey_backfills.py. The suggester
identified 3 orphan bib entries whose arXiv ids already appear in this
file's References sections — the paper was cited, just without the
canonical [@bibkey] link. Appended the bibkey to the end of each
reference line:

  copula_sensitivity         -> [@balgi2025sensitivity]     (arXiv:2508.08752)
  survival_sensitivity       -> [@hu2025nonparametric]      (arXiv:2511.01412)
  calibrate_confounding_strength -> [@baitairian2025calibrating] (arXiv:2510.16560)

Only the clean end-of-reference lines were touched; the mid-sentence
summary mentions at the top of the module docstring were deliberately
left alone (StatsPAI Phase 2 convention: bibkey annotations live in
the References block, not in prose).

Pure metadata — no function signatures, numerics, or algorithm paths
changed.

Verification
------------
  coverage:        64.0% -> 64.6%        (315 -> 318 cited keys)
  orphans:         177   -> 174          (3 keys fully consumed)
  candidate lines: 65    -> 59           (suggester re-run)
  candidate keys:  21    -> 18

  Gate 0  pytest 81 tests                                  pass
  Gate 1  audit_bib_duplicates --strict   0/0/0            pass
  Gate 2  audit_bib_coverage  --strict-dangling  0 dangling pass
  Gate 3  audit_citations     --strict    316/0/0          pass

Next bibkey backfills (18 keys / 59 lines remain). Tool output is
reproducible with:
  python3 tools/suggest_bibkey_backfills.py

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

Author: brycewang-stanford

src/statspai/robustness/sensitivity_frontier.py

@@ -107,7 +107,7 @@ def copula_sensitivity(
 
     References
     ----------
-    Balgi, Braun, Peña & Daoud (arXiv:2508.08752, 2025).
+    Balgi, Braun, Peña & Daoud (arXiv:2508.08752, 2025). [@balgi2025sensitivity]
     """
     if rho_grid is None:
         rho_grid = np.linspace(-0.5, 0.5, 21)
@@ -185,7 +185,7 @@ def survival_sensitivity(
 
     References
     ----------
-    Hu & Westling (arXiv:2511.01412, 2025).
+    Hu & Westling (arXiv:2511.01412, 2025). [@hu2025nonparametric]
     """
     if se_log_hr <= 0:
         raise ValueError("`se_log_hr` must be > 0.")
@@ -267,7 +267,7 @@ def calibrate_confounding_strength(
 
     References
     ----------
-    Baitairian et al. (arXiv:2510.16560, 2025).
+    Baitairian et al. (arXiv:2510.16560, 2025). [@baitairian2025calibrating]
     Cinelli & Hazlett (JRSS-B 2020).
     """
     if not (0 < observed_r2_outcome < 1):