fix(cli): propagate --repo/--path to search/chat subparsers, fix --ba…

docs/guides/email.mdx

@@ -127,16 +127,16 @@ Each row carries inline action buttons:
 
 If you haven't connected Google yet, the agent surfaces a one-click **Connect Google** button inline in the chat — no need to navigate to Settings → Connections manually.
 
-### In-session preferences (in-memory, wiped on restart)
+### Cross-session preferences (persisted via memory store)
 
-Tell the agent how you want classification to behave for this session:
+Tell the agent how you want classification to behave:
 
-- *"Treat boss@company.com as urgent"* → calls `set_priority_sender`. That sender bypasses the heuristic and lands in **Urgent** for the rest of the session.
+- *"Treat boss@company.com as urgent"* → calls `set_priority_sender`. That sender bypasses the heuristic and lands in **Urgent** across all future sessions.
 - *"Treat newsletter@stripe.com as low priority"* → calls `set_low_priority_sender`. That sender lands in **Suggested archives**.
 - *"Default informational mail to archive"* → calls `set_category_default("informational", "archive")`. Informational items lift into **Suggested archives** until you reset.
 - *"Clear my preferences"* → calls `clear_session_preferences`.
 
-Preferences are stored in process memory only — restarting the agent (or quitting Agent UI) wipes them. This is deliberate: the goal is to prove the value of session-scoped learning before we wire up persistent memory. Once persistent memory ships, the same tools will write through to it without changing this surface.
+Preferences persist across agent restarts via the memory store (`email:preferences`). They fall back to session-only (lost on restart) when `GAIA_MEMORY_DISABLED=1` or when the embedding backend is unreachable at startup. Incognito sessions never write preferences to persistent storage.
 
 ### Behavioral learning (auto-promotion)
 
@@ -148,7 +148,7 @@ Senders you reply to quickly are automatically promoted to priority on the next
 
 `list_inbox`, `get_message`, `get_thread`, `search_messages`, `list_labels`, `triage_inbox`, `pre_scan_inbox`
 
-### Session preferences (in-memory; wiped on agent restart)
+### Preferences (persisted across restarts via memory store)
 
 `set_priority_sender`, `set_low_priority_sender`, `set_category_default`, `clear_session_preferences`
 

hub/agents/python/code/gaia_agent_code/cli.py

@@ -284,7 +284,7 @@ def _build_index_parser():
     parser.add_argument(
         "--base-url",
         default=None,
-        help="Lemonade server URL (default: http://localhost:8000/api/v1)",
+        help="Lemonade server URL (default: http://localhost:13305/api/v1)",
     )
     parser.add_argument(
         "--no-lemonade-check",
@@ -308,10 +308,30 @@ def _build_index_parser():
         default=10,
         help="Number of results to return (default: 10)",
     )
+    # FIX 1: use SUPPRESS so the parent's default="." is preserved when --repo is omitted
+    search_p.add_argument(
+        "--repo",
+        default=argparse.SUPPRESS,
+        help="Path to repository root (overrides parent --repo)",
+    )
 
     sub.add_parser("status", help="Show index status")
     sub.add_parser("clear", help="Clear the index")
-    sub.add_parser("chat", help="Interactive code Q&A (CodeAgent + code_index tools)")
+
+    chat_p = sub.add_parser("chat", help="Interactive code Q&A (CodeAgent + code_index tools)")
+    # FIX 2: use SUPPRESS so the parent's default="." is preserved when --repo is omitted
+    chat_p.add_argument(
+        "--repo",
+        default=argparse.SUPPRESS,
+        help="Path to repository root (overrides parent --repo)",
+    )
+    # FIX 3: dest="repo" makes --path a true alias for --repo, no changes needed in cmd_index
+    chat_p.add_argument(
+        "--path",
+        dest="repo",
+        default=argparse.SUPPRESS,
+        help="Alias for --repo (path to repository root)",
+    )
 
     return parser
 
@@ -509,7 +529,7 @@ def main():
     parser.add_argument(
         "--path",
         "-p",
-        type=str,
+        type=str,   
         default=None,
         help="Project directory path. Creates directory if it doesn't exist.",
     )
@@ -606,4 +626,4 @@ def main():
 
 
 if __name__ == "__main__":
-    sys.exit(main())
+    sys.exit(main())

hub/agents/python/email/gaia_agent_email/tools/profile_tools.py

@@ -19,12 +19,17 @@
 
 Bounded and idempotent by design — ``_record_interaction`` always does an
 upsert into the single per-sender record, so the record count is O(senders)
+<<<<<<< HEAD
+regardless of how many emails arrive. #1290 can reuse ``_read_interactions``
+directly to build richer views.
+=======
 regardless of how many emails arrive.
 
 Reply behavior is tracked separately in reply records (entity
 ``email:reply:<sender>``). ``_evaluate_promotions()`` reads these to identify
 senders with consistently fast replies and returns them for promotion to
 priority senders. Called on-demand at triage time — no background thread.
+>>>>>>> upstream/main
 
 Tools registered:
 
@@ -52,6 +57,8 @@
 _INTERACTION_DOMAIN = "email_agent_interactions"
 _INTERACTION_CATEGORY = "interaction"
 
+<<<<<<< HEAD
+=======
 # Stable entity prefix for per-sender REPLY behavior records.
 # Entity = f"{_REPLY_ENTITY_PREFIX}{sender_email}"
 _REPLY_ENTITY_PREFIX = "email:reply:"
@@ -79,6 +86,7 @@
 # 50k senders is far beyond any realistic mailbox.
 _MAX_REPLY_RECORDS: int = 50000
 
+>>>>>>> upstream/main
 # Sanity ceiling on how many distinct-sender interaction records we read in
 # one profiling pass. This is NOT a silent truncation cap: if a real inbox
 # ever has more distinct senders than this, ``_read_interactions`` logs a
@@ -101,13 +109,25 @@ def _now_iso() -> str:
 
 
 def _dominant_category(category_counts: Dict[str, int]) -> str:
+<<<<<<< HEAD
+    """Return the category with the highest count; ties broken by last alphabetically."""
+=======
     """Return the category with the highest count (ties: lexicographic order)."""
+>>>>>>> upstream/main
     if not category_counts:
         return ""
     return max(category_counts, key=lambda cat: (category_counts[cat], cat))
 
 
 class ProfileToolsMixin:
+<<<<<<< HEAD
+    """Mixin that registers inbox-profiling tools.
+
+    State-free at construction time — reads ``self._memory_store`` via
+    a closure captured when ``_register_profile_tools()`` is called.
+    """
+
+=======
     """Mixin that registers inbox-profiling tools and behavioral-learning helpers.
 
     State-free at construction time — reads ``self._memory_store`` via
@@ -236,6 +256,7 @@ def _evaluate_promotions(self) -> List[str]:
                 )
         return qualified
 
+>>>>>>> upstream/main
     def _record_interaction(self, sender: str, category: str) -> None:
         """Update the rolling interaction record for *sender*.