feat: live progress updates during orchestrator tool loop

The thinking bubble now shows real-time status instead of a static spinner:
  ⚡ Round 1 — thinking…
  ⚡ Round 1 — web_search
  ⚡ Round 2 — thinking…
  ⚡ Generating response…

Implementation: async on_progress callback passed from _run_job into both
orchestrators (_run_from_messages / _run_from_contents). Callback writes to
job["progress"] under the jobs lock; poll responses include the field;
app.js displays it in the thinking bubble when present.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

cortex/openai_orchestrator.py

@@ -52,6 +52,7 @@ async def run(
     max_risk: str | None = None,
     risk_whitelist: list[str] | None = None,
     risk_blacklist: list[str] | None = None,
+    on_progress=None,  # async (str) -> None; called with live status updates
 ) -> OrchestratorResult:
     """
     Run a tool-enabled task using an OpenAI-compatible API.
@@ -117,6 +118,7 @@ async def run(
         confirm_allow=_confirm_allow,
         confirm_deny=_confirm_deny,
         starting_round=0,
+        on_progress=on_progress,
     )
 
     if checkpoint:
@@ -307,6 +309,7 @@ async def _run_from_messages(
     confirm_deny: frozenset,
     starting_round: int = 0,
     tool_list: list[str] | None = None,
+    on_progress=None,
 ) -> tuple[str, OrchestrateCheckpoint | None]:
     """
     Run the OpenAI ReAct loop from the current messages state.
@@ -323,6 +326,8 @@ async def _run_from_messages(
         est = _estimate_tokens(messages)
         logger.info("OpenAI orchestrator round %d / %d  model=%s  ~%d tokens",
                     round_num + 1, effective_limit, model_name, est)
+        if on_progress:
+            await on_progress(f"Round {round_num + 1} — thinking…")
 
         call_kwargs: dict = {"model": model_name, "messages": messages}
         if active_tools:
@@ -373,6 +378,8 @@ async def _run_from_messages(
                     pending_tools.append({"name": name, "args": args_parsed, "tool_call_id": tc.id})
                     logger.info("Tool %s blocked — confirmation required", name)
                 else:
+                    if on_progress:
+                        await on_progress(f"Round {round_num + 1} — {name}")
                     result_str = await _execute_tool(name, tc.function.arguments, user_role, tool_list)
                     logger.info("Tool %s → %d chars", name, len(result_str))
                     executed_results.append({"name": name, "args": args_parsed, "result": result_str, "tool_call_id": tc.id})
@@ -415,6 +422,8 @@ async def _run_from_messages(
                 return final_response, checkpoint
 
         else:
+            if on_progress:
+                await on_progress("Generating response…")
             final_response = msg.content or ""
             logger.info(
                 "OpenAI orchestrator done after %d round(s). Tools used: %d",