feat: add official Codex skill and harden task execution

Install and update artemis-use with task diagnostics and plugin guidance. Preserve cross-device file receipts, review canonical reports, guard stage transitions and reduce small history summaries. Improve exploration lifecycle, request recovery and browser live connections.

tests/unit/test_llm_gateway.py

@@ -319,7 +319,8 @@ async def test_mid_stream_failure_records_stream_reset_payload(monkeypatch):
     assert payload["action"] == "discard"
     assert payload["reason"] == "mid_stream_failure"
     assert "stream_exec_id" in payload
-    assert "lower API priority" in payload["message"]
+    assert "未完成的输出已丢弃" in payload["message"]
+    assert "lower API priority" not in payload["message"]
 
 
 @pytest.mark.asyncio