diff --git a/README.md b/README.md
index 72046e7..b5699bb 100644
--- a/README.md
+++ b/README.md
@@ -45,3 +45,8 @@ This repository now has a dedicated runtime layer and CLI diagnostics. The next
 • services/
 • runtime/
 • domain logic
+
+The first split is already in place for:
+
+• smart-executor -> commands.smart_executor + services.smart_executor_service
+• precheck -> commands.precheck + services.precheck_service
diff --git a/src/coinhunter/cli.py b/src/coinhunter/cli.py
index 0a44ca4..451ba03 100755
--- a/src/coinhunter/cli.py
+++ b/src/coinhunter/cli.py
@@ -15,11 +15,11 @@ MODULE_MAP = {
     "init": "init_user_state",
     "market-probe": "market_probe",
     "paths": "paths",
-    "precheck": "precheck",
+    "precheck": "commands.precheck",
     "review-context": "review_context",
     "review-engine": "review_engine",
     "rotate-external-gate-log": "rotate_external_gate_log",
-    "smart-executor": "smart_executor",
+    "smart-executor": "commands.smart_executor",
     "auto-trader": "auto_trader",
 }
 
diff --git a/src/coinhunter/commands/__init__.py b/src/coinhunter/commands/__init__.py
new file mode 100644
index 0000000..7838ce4
--- /dev/null
+++ b/src/coinhunter/commands/__init__.py
@@ -0,0 +1 @@
+"""CLI command adapters for CoinHunter."""
diff --git a/src/coinhunter/commands/precheck.py b/src/coinhunter/commands/precheck.py
new file mode 100644
index 0000000..8f38d99
--- /dev/null
+++ b/src/coinhunter/commands/precheck.py
@@ -0,0 +1,15 @@
+"""CLI adapter for precheck."""
+
+from __future__ import annotations
+
+import sys
+
+from ..services.precheck_service import run
+
+
+def main() -> int:
+    return run(sys.argv[1:])
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())
diff --git a/src/coinhunter/commands/smart_executor.py b/src/coinhunter/commands/smart_executor.py
new file mode 100644
index 0000000..3ab77b8
--- /dev/null
+++ b/src/coinhunter/commands/smart_executor.py
@@ -0,0 +1,15 @@
+"""CLI adapter for smart executor."""
+
+from __future__ import annotations
+
+import sys
+
+from ..services.smart_executor_service import run
+
+
+def main() -> int:
+    return run(sys.argv[1:])
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())
diff --git a/src/coinhunter/precheck.py b/src/coinhunter/precheck.py
index 548e9a7..3f68a06 100755
--- a/src/coinhunter/precheck.py
+++ b/src/coinhunter/precheck.py
@@ -924,35 +924,9 @@ def ack_analysis(note: str = ""):
 
 
 def main():
-    if len(sys.argv) > 1 and sys.argv[1] == "--ack":
-        ack_analysis(" ".join(sys.argv[2:]).strip())
-        return
-    if len(sys.argv) > 1 and sys.argv[1] == "--mark-run-requested":
-        mark_run_requested(" ".join(sys.argv[2:]).strip())
-        return
+    from .services.precheck_service import run
 
-    try:
-        state = sanitize_state_for_stale_triggers(load_state())
-        snapshot = build_snapshot()
-        analysis = analyze_trigger(snapshot, state)
-        save_state(update_state_after_observation(state, snapshot, analysis))
-        print(json.dumps(analysis, ensure_ascii=False, indent=2))
-    except Exception as e:
-        failure = {
-            "generated_at": utc_iso(),
-            "status": "deep_analysis_required",
-            "should_analyze": True,
-            "pending_trigger": True,
-            "cooldown_active": False,
-            "reasons": ["precheck-error"],
-            "hard_reasons": ["precheck-error"],
-            "soft_reasons": [],
-            "soft_score": 0,
-            "details": [str(e)],
-            "compact_summary": f"预检查失败，转入深度分析兜底: {e}",
-        }
-        print(json.dumps(failure, ensure_ascii=False, indent=2))
-        sys.exit(0)
+    return run(sys.argv[1:])
 
 
 if __name__ == "__main__":
diff --git a/src/coinhunter/services/__init__.py b/src/coinhunter/services/__init__.py
new file mode 100644
index 0000000..63957c1
--- /dev/null
+++ b/src/coinhunter/services/__init__.py
@@ -0,0 +1 @@
+"""Application services for CoinHunter."""
diff --git a/src/coinhunter/services/precheck_service.py b/src/coinhunter/services/precheck_service.py
new file mode 100644
index 0000000..b5a2c36
--- /dev/null
+++ b/src/coinhunter/services/precheck_service.py
@@ -0,0 +1,43 @@
+"""Service entrypoint for precheck workflows."""
+
+from __future__ import annotations
+
+import json
+import sys
+
+from .. import precheck as precheck_module
+
+
+def run(argv: list[str] | None = None) -> int:
+    argv = list(sys.argv[1:] if argv is None else argv)
+
+    if argv and argv[0] == "--ack":
+        precheck_module.ack_analysis(" ".join(argv[1:]).strip())
+        return 0
+    if argv and argv[0] == "--mark-run-requested":
+        precheck_module.mark_run_requested(" ".join(argv[1:]).strip())
+        return 0
+
+    try:
+        state = precheck_module.sanitize_state_for_stale_triggers(precheck_module.load_state())
+        snapshot = precheck_module.build_snapshot()
+        analysis = precheck_module.analyze_trigger(snapshot, state)
+        precheck_module.save_state(precheck_module.update_state_after_observation(state, snapshot, analysis))
+        print(json.dumps(analysis, ensure_ascii=False, indent=2))
+        return 0
+    except Exception as exc:
+        failure = {
+            "generated_at": precheck_module.utc_iso(),
+            "status": "deep_analysis_required",
+            "should_analyze": True,
+            "pending_trigger": True,
+            "cooldown_active": False,
+            "reasons": ["precheck-error"],
+            "hard_reasons": ["precheck-error"],
+            "soft_reasons": [],
+            "soft_score": 0,
+            "details": [str(exc)],
+            "compact_summary": f"预检查失败，转入深度分析兜底: {exc}",
+        }
+        print(json.dumps(failure, ensure_ascii=False, indent=2))
+        return 0
diff --git a/src/coinhunter/services/smart_executor_service.py b/src/coinhunter/services/smart_executor_service.py
new file mode 100644
index 0000000..4d122bd
--- /dev/null
+++ b/src/coinhunter/services/smart_executor_service.py
@@ -0,0 +1,111 @@
+"""Service entrypoint for smart executor workflows."""
+
+from __future__ import annotations
+
+import os
+import sys
+
+from .. import smart_executor as smart_executor_module
+
+
+def run(argv: list[str] | None = None) -> int:
+    argv = list(sys.argv[1:] if argv is None else argv)
+    args, normalized_argv = smart_executor_module.parse_cli_args(argv)
+    action = args.command.replace("-", "_")
+    argv_tail = smart_executor_module.cli_action_args(args, action)
+    decision_id = (
+        args.decision_id
+        or os.getenv("DECISION_ID")
+        or smart_executor_module.default_decision_id(action, normalized_argv)
+    )
+    if args.dry_run:
+        smart_executor_module.DRY_RUN = True
+
+    previous = smart_executor_module.get_execution_state(decision_id)
+    read_only_action = action in {"balance", "balances", "status"}
+    if previous and previous.get("status") == "success" and not read_only_action:
+        smart_executor_module.log(f"⚠️ decision_id={decision_id} 已执行成功，跳过重复执行")
+        return 0
+
+    try:
+        ex = smart_executor_module.get_exchange()
+        if read_only_action:
+            if action in {"balance", "balances"}:
+                smart_executor_module.command_balances(ex)
+            else:
+                smart_executor_module.command_status(ex)
+            return 0
+
+        decision_context = smart_executor_module.build_decision_context(ex, action, argv_tail, decision_id)
+        if args.analysis:
+            decision_context["analysis"] = args.analysis
+        elif os.getenv("DECISION_ANALYSIS"):
+            decision_context["analysis"] = os.getenv("DECISION_ANALYSIS")
+        if args.reasoning:
+            decision_context["reasoning"] = args.reasoning
+        elif os.getenv("DECISION_REASONING"):
+            decision_context["reasoning"] = os.getenv("DECISION_REASONING")
+
+        smart_executor_module.record_execution_state(
+            decision_id,
+            {"status": "pending", "started_at": smart_executor_module.bj_now_iso(), "action": action, "args": argv_tail},
+        )
+
+        if action == "sell_all":
+            result = smart_executor_module.action_sell_all(ex, args.symbol, decision_id, decision_context)
+        elif action == "buy":
+            result = smart_executor_module.action_buy(ex, args.symbol, float(args.amount_usdt), decision_id, decision_context)
+        elif action == "rebalance":
+            result = smart_executor_module.action_rebalance(ex, args.from_symbol, args.to_symbol, decision_id, decision_context)
+        elif action == "hold":
+            balances = smart_executor_module.fetch_balances(ex)
+            positions = smart_executor_module.load_positions()
+            market_snapshot = smart_executor_module.build_market_snapshot(ex)
+            smart_executor_module.log_decision(
+                {
+                    **decision_context,
+                    "balances_after": balances,
+                    "positions_after": positions,
+                    "market_snapshot": market_snapshot,
+                    "analysis": decision_context.get("analysis", "hold"),
+                    "reasoning": decision_context.get("reasoning", "hold"),
+                    "execution_result": {"status": "hold"},
+                }
+            )
+            smart_executor_module.log("😴 决策: 持续持有，无操作")
+            result = {"status": "hold"}
+        else:
+            raise RuntimeError(f"未知动作: {action}；请运行 --help 查看正确 CLI 用法")
+
+        smart_executor_module.record_execution_state(
+            decision_id,
+            {
+                "status": "success",
+                "finished_at": smart_executor_module.bj_now_iso(),
+                "action": action,
+                "args": argv_tail,
+                "result": result,
+            },
+        )
+        smart_executor_module.log(f"✅ 执行完成 decision_id={decision_id}")
+        return 0
+    except Exception as exc:
+        smart_executor_module.record_execution_state(
+            decision_id,
+            {
+                "status": "failed",
+                "finished_at": smart_executor_module.bj_now_iso(),
+                "action": action,
+                "args": argv_tail,
+                "error": str(exc),
+            },
+        )
+        smart_executor_module.log_error(
+            "smart_executor",
+            exc,
+            decision_id=decision_id,
+            action=action,
+            args=argv_tail,
+        )
+        smart_executor_module.log(f"❌ 执行失败: {exc}")
+        return 1
diff --git a/src/coinhunter/smart_executor.py b/src/coinhunter/smart_executor.py
index 47815c6..954bb02 100755
--- a/src/coinhunter/smart_executor.py
+++ b/src/coinhunter/smart_executor.py
@@ -538,72 +538,9 @@ def cli_action_args(args, action: str) -> list[str]:
 
 
 def main():
-    global DRY_RUN
-    args, normalized_argv = parse_cli_args(sys.argv[1:])
-    action = args.command.replace("-", "_")
-    argv_tail = cli_action_args(args, action)
-    decision_id = args.decision_id or os.getenv("DECISION_ID") or default_decision_id(action, normalized_argv)
-    if args.dry_run:
-        DRY_RUN = True
+    from .services.smart_executor_service import run
 
-    previous = get_execution_state(decision_id)
-    read_only_action = action in {"balance", "balances", "status"}
-    if previous and previous.get("status") == "success" and not read_only_action:
-        log(f"⚠️ decision_id={decision_id} 已执行成功，跳过重复执行")
-        return
-
-    try:
-        ex = get_exchange()
-        if read_only_action:
-            if action in {"balance", "balances"}:
-                command_balances(ex)
-            else:
-                command_status(ex)
-            return
-
-        decision_context = build_decision_context(ex, action, argv_tail, decision_id)
-        if args.analysis:
-            decision_context["analysis"] = args.analysis
-        elif os.getenv("DECISION_ANALYSIS"):
-            decision_context["analysis"] = os.getenv("DECISION_ANALYSIS")
-        if args.reasoning:
-            decision_context["reasoning"] = args.reasoning
-        elif os.getenv("DECISION_REASONING"):
-            decision_context["reasoning"] = os.getenv("DECISION_REASONING")
-
-        record_execution_state(decision_id, {"status": "pending", "started_at": bj_now_iso(), "action": action, "args": argv_tail})
-
-        if action == "sell_all":
-            result = action_sell_all(ex, args.symbol, decision_id, decision_context)
-        elif action == "buy":
-            result = action_buy(ex, args.symbol, float(args.amount_usdt), decision_id, decision_context)
-        elif action == "rebalance":
-            result = action_rebalance(ex, args.from_symbol, args.to_symbol, decision_id, decision_context)
-        elif action == "hold":
-            balances = fetch_balances(ex)
-            positions = load_positions()
-            market_snapshot = build_market_snapshot(ex)
-            log_decision({
-                **decision_context,
-                "balances_after": balances,
-                "positions_after": positions,
-                "market_snapshot": market_snapshot,
-                "analysis": decision_context.get("analysis", "hold"),
-                "reasoning": decision_context.get("reasoning", "hold"),
-                "execution_result": {"status": "hold"},
-            })
-            log("😴 决策: 持续持有，无操作")
-            result = {"status": "hold"}
-        else:
-            raise RuntimeError(f"未知动作: {action}；请运行 --help 查看正确 CLI 用法")
-
-        record_execution_state(decision_id, {"status": "success", "finished_at": bj_now_iso(), "action": action, "args": argv_tail, "result": result})
-        log(f"✅ 执行完成 decision_id={decision_id}")
-    except Exception as e:
-        record_execution_state(decision_id, {"status": "failed", "finished_at": bj_now_iso(), "action": action, "args": argv_tail, "error": str(e)})
-        log_error("smart_executor", e, decision_id=decision_id, action=action, args=argv_tail)
-        log(f"❌ 执行失败: {e}")
-        sys.exit(1)
+    return run(sys.argv[1:])
 
 
 if __name__ == "__main__":