refactor: bind gateway anti-growth guard to locals, trim overlong comment

Simplify-code pass: gateway/run.py called estimate_messages_tokens_rough
6x on the same data in the anti-growth guard (condition + warning f-string).
Bind to _hyg_in_toks/_hyg_out_toks locals like the conversation_compression.py
guard already does. Also trim the comment from 10 lines to 4 (keep the WHY,
drop the WHAT) and remove an extra blank line before TestCompactedTurnsStaySearchable.

gateway/run.py

@@ -19031,28 +19031,20 @@ class GatewayRunner(GatewayAuthorizationMixin, GatewayKanbanWatchersMixin, Gatew
                                     _hyg_in_place = bool(
                                         getattr(_hyg_agent, "_last_compaction_in_place", False)
                                     )
-                                    # Anti-regression guard: never persist a
-                                    # compression that did not shrink the
-                                    # transcript. A summary that is larger than
-                                    # the middle it replaces would GROW the
-                                    # session instead of reclaiming it (observed:
-                                    # 427K -> 598K). Compare like-for-like (both
-                                    # rough estimates) so an "actual vs
-                                    # estimate" measurement mismatch can't
-                                    # produce a false verdict. On growth, keep
-                                    # the original transcript untouched.
-                                    if (
-                                        _hyg_rotated
-                                    ) and estimate_messages_tokens_rough(
-                                        _compressed
-                                    ) > estimate_messages_tokens_rough(history):
+                                    # Anti-growth guard: refuse a compression
+                                    # that did not shrink the transcript
+                                    # (observed: 427K -> 598K). Compare
+                                    # like-for-like rough estimates.
+                                    _hyg_in_toks = estimate_messages_tokens_rough(history)
+                                    _hyg_out_toks = estimate_messages_tokens_rough(_compressed)
+                                    if _hyg_rotated and _hyg_out_toks > _hyg_in_toks:
                                         logger.warning(
                                             "Gateway hygiene compression for session %s "
                                             "would grow transcript (~%s -> ~%s tokens); "
                                             "keeping the original transcript unchanged",
                                             session_entry.session_id,
-                                            f"{estimate_messages_tokens_rough(history):,}",
-                                            f"{estimate_messages_tokens_rough(_compressed):,}",
+                                            f"{_hyg_in_toks:,}",
+                                            f"{_hyg_out_toks:,}",
                                         )
                                         _hyg_rotated = False
                                         _compressed = history

tests/run_agent/test_in_place_compaction.py

@@ -320,7 +320,6 @@ class TestInPlaceAntiGrowthGuard:
             ]
 
 
-
 class TestCompactedTurnsStaySearchable:
     """Teknium's review hinges on the pre-compaction transcript staying
     DISCOVERABLE after in-place compaction. Compaction-archived rows