refactor: address high-priority debt and publish to PyPI

- Fix TOCTOU race conditions by wrapping read-modify-write cycles under single-file locks in execution_state, portfolio_service, precheck_state, state_manager, and precheck_service. - Add missing test coverage (96 tests total): - test_review_service.py (15 tests) - test_check_api.py (6 tests) - test_external_gate.py main branches (+10 tests) - test_trade_execution.py new commands (+8 tests) - Unify all agent-consumed JSON messages to English. - Config-ize hardcoded values (volume filter, schema_version) via get_user_config with sensible defaults. - Add 1-hour TTL to exchange cache with force_new override. - Add ruff and mypy to dev dependencies; fix all type errors. - Add __all__ declarations to 11 service modules. - Sync README with new commands, config tuning docs, and PyPI badge. - Publish package as coinhunter==1.0.0 on PyPI with MIT license. - Deprecate coinhunter-cli==1.0.1 with runtime warning. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-04-16 01:21:27 +08:00
parent 01bb54dee5
commit 62c40a9776
53 changed files with 2338 additions and 671 deletions
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -6,7 +6,7 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co

 - **Editable install:** `pip install -e .`
 - **Run the CLI locally:** `python -m coinhunter --help`
- **Install for end users:** `./scripts/install_local.sh` (creates `~/.local/bin/coinhunter`)
+- **Install for end users:** `./scripts/install_local.sh` (standard `pip install -e .` wrapper)
 - **Tests:** There is no test suite yet. The README lists next priorities as adding pytest coverage for runtime paths, state manager, and trigger analyzer.
 - **Lint / type-check:** Not configured yet.

@@ -14,7 +14,7 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co

 `src/coinhunter/cli.py` is the single entrypoint. It resolves aliases to canonical commands, maps canonical commands to Python modules via `MODULE_MAP`, then imports the module and calls `module.main()` after mutating `sys.argv` to match the display name.

-Active commands live in `src/coinhunter/commands/` and are thin adapters that delegate to `src/coinhunter/services/`. A few legacy commands (`review_context.py`, `review_engine.py`) still live at the `src/coinhunter/` root and are imported directly by `cli.py`. Root-level backward-compat facades (e.g., `precheck.py`, `smart_executor.py`) re-export the moved commands.
+Active commands live in `src/coinhunter/commands/` and are thin adapters that delegate to `src/coinhunter/services/`. Root-level backward-compat facades (e.g., `precheck.py`, `smart_executor.py`, `review_engine.py`, `review_context.py`) re-export the moved commands.

 ## Architecture

@@ -28,7 +28,7 @@ Active commands live in `src/coinhunter/commands/` and are thin adapters that de

 ### Runtime and environment

-`runtime.py` defines `RuntimePaths` and `get_runtime_paths()`. User data lives in `~/.coinhunter/` by default (override with `COINHUNTER_HOME`). Credentials are loaded from `~/.hermes/.env` by default (override with `COINHUNTER_ENV_FILE`). Many modules eagerly instantiate `PATHS = get_runtime_paths()` at import time.
+`runtime.py` defines `RuntimePaths` and `get_runtime_paths()`. User data lives in `~/.coinhunter/` by default (override with `COINHUNTER_HOME`). Credentials are loaded from `~/.hermes/.env` by default (override with `COINHUNTER_ENV_FILE`). Modules should call `get_runtime_paths()` at function scope rather than eagerly at import time.

 ### Smart executor (`exec`)

@@ -56,8 +56,8 @@ State is stored in `~/.coinhunter/state/precheck_state.json`.

 ### Review commands

- `review N` (`review_context.py`) — generates review context for the last N hours.
- `recap N` (`review_engine.py`) — generates a full review report by reading JSONL decision/trade/error logs, computing PnL estimates, missed opportunities, and saving a report to `~/.coinhunter/reviews/`.
+- `review N` (`commands/review_context.py` → `services/review_service.py`) — generates review context for the last N hours.
+- `recap N` (`commands/review_engine.py` → `services/review_service.py`) — generates a full review report by reading JSONL decision/trade/error logs, computing PnL estimates, missed opportunities, and saving a report to `~/.coinhunter/reviews/`.

 ### Logging model

--- a/21
+++ b/21
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Tacit Lab
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
--- a/README.md
+++ b/README.md
@@ -23,17 +23,19 @@
 </p>

 <p align="center">
+  <a href="https://pypi.org/project/coinhunter/"><img src="https://img.shields.io/pypi/v/coinhunter?style=flat-square&color=3776ab" /></a>
  <a href="#"><img src="https://img.shields.io/badge/python-3.10%2B-3776ab?style=flat-square&logo=python&logoColor=white" /></a>
  <a href="#"><img src="https://img.shields.io/badge/status-active%20development-22c55e?style=flat-square" /></a>
  <a href="#"><img src="https://img.shields.io/badge/architecture-cli%20%2B%20commands%20%2B%20services-8b5cf6?style=flat-square" /></a>
-  <a href="#"><img src="https://img.shields.io/badge/tests-needed-ef4444?style=flat-square" /></a>
 </p>

 ---

 ## What is this?

-`coinhunter-cli` is the **executable tooling layer** for CoinHunter — an installable Python CLI that handles trading operations, market probes, precheck orchestration, and review workflows.
+`coinhunter` is the **executable tooling layer** for CoinHunter — an installable Python CLI that handles trading operations, market probes, precheck orchestration, and review workflows.
+
+> **Note:** The old package name `coinhunter-cli` is deprecated. Please install `coinhunter` going forward.

 | Layer | Responsibility | Location |
 |-------|----------------|----------|
@@ -74,7 +76,8 @@ src/coinhunter/
 │   ├── smart_executor_parser.py
 │   ├── execution_state.py
 │   ├── precheck_service.py
-│   ├── precheck_constants.py     # thresholds & paths
+│   ├── review_service.py         # review generation logic
+│   ├── precheck_constants.py     # thresholds
 │   ├── time_utils.py             # UTC/local time helpers
 │   ├── data_utils.py             # json, hash, float, symbol norm
 │   ├── state_manager.py          # state load/save/sanitize
@@ -96,21 +99,13 @@ src/coinhunter/

 ## Installation

-### Quick user install
+### From PyPI (recommended)

 ```bash
-./scripts/install_local.sh
+pip install coinhunter
 ```

-Creates:
- Virtualenv: `~/.local/share/coinhunter-cli/venv`
- Launcher: `~/.local/bin/coinhunter`
-
-### Editable dev install
-
-```bash
-pip install -e .
-```
+This installs the latest stable release and creates the `coinhunter` console script entry point.

 Verify:

@@ -119,6 +114,22 @@ coinhunter --help
 coinhunter --version
 ```

+### Development install (editable)
+
+If you're working on this repo locally:
+
+```bash
+pip install -e ".[dev]"
+```
+
+Or use the convenience script:
+
+```bash
+./scripts/install_local.sh
+```
+
+A thin wrapper that runs `pip install -e .` and verifies the entrypoint is on your PATH.
+
 ---

 ## Command Reference
@@ -136,6 +147,9 @@ coinhunter exec hold         # record a HOLD decision
 coinhunter exec buy ENJUSDT 50   # buy $50 of ENJUSDT
 coinhunter exec flat ENJUSDT     # sell entire ENJUSDT position
 coinhunter exec rotate PEPEUSDT ETHUSDT  # rotate exposure
+coinhunter exec orders       # list open spot orders
+coinhunter exec order-status ENJUSDT 123456  # check specific order
+coinhunter exec cancel ENJUSDT 123456        # cancel an open order
 coinhunter gate              # external gate orchestration
 coinhunter review 12         # generate review context (last 12h)
 coinhunter recap 12          # generate review report (last 12h)
@@ -208,6 +222,45 @@ Run the external gate:
 coinhunter gate
 ```

+The gate reads `trigger_command` from `~/.coinhunter/config.json` under `external_gate`.
+- By default, no external trigger is configured — gate runs precheck and marks state, then exits cleanly.
+- Set `trigger_command` to a command list to integrate with your own scheduler:
+
+```json
+{
+  "external_gate": {
+    "trigger_command": ["hermes", "cron", "run", "JOB_ID"]
+  }
+}
+```
+
+- Set to `null` or `[]` to explicitly disable the external trigger.
+
+### Dynamic tuning via `config.json`
+
+You can override internal defaults without editing code by adding keys to `~/.coinhunter/config.json`:
+
+```json
+{
+  "external_gate": {
+    "trigger_command": ["hermes", "cron", "run", "JOB_ID"]
+  },
+  "exchange": {
+    "min_quote_volume": 200000,
+    "cache_ttl_seconds": 3600
+  },
+  "logging": {
+    "schema_version": 2
+  }
+}
+```
+
+| Key | Default | Effect |
+|-----|---------|--------|
+| `exchange.min_quote_volume` | `200000` | Minimum 24h quote volume for a symbol to appear in market snapshots |
+| `exchange.cache_ttl_seconds` | `3600` | How long the ccxt exchange instance (and `load_markets()` result) is cached |
+| `logging.schema_version` | `2` | Schema version stamped on every JSONL log entry |
+
 ---

 ## Runtime Model
@@ -248,14 +301,17 @@ The codebase is actively maintained and refactored in small, safe steps.
 - ✅ Extracted `smart-executor` into `commands/` + `services/`
 - ✅ Extracted `precheck` into 9 focused service modules
 - ✅ Migrated all active command modules into `commands/`
+- ✅ Extracted `review_engine.py` core logic into `services/review_service.py`
+- ✅ Removed eager `PATHS` instantiation across services and commands
+- ✅ Fixed `smart_executor.py` lazy-loading facade
+- ✅ Standardized install to use `pip install -e .`
+- ✅ Made `external_gate` trigger_command configurable (no longer hardcodes hermes)
 - ✅ Removed dead `auto-trader` command
 - ✅ Backward-compatible root facades preserved

 **Next priorities:**
- 🧪 Add pytest coverage for runtime paths, state manager, and trigger analyzer
- 🔧 Extract `review_engine.py` core logic into `services/`
- 🔧 Unify output contract (JSON-first with `--pretty` option)
 - 🔧 Add basic CI (lint + compileall + pytest)
+- 🔧 Unify output contract (JSON-first with `--pretty` option)

 ---

--- a/pyproject.toml
+++ b/pyproject.toml
@@ -3,10 +3,11 @@ requires = ["setuptools>=68", "wheel"]
 build-backend = "setuptools.build_meta"

 [project]
-name = "coinhunter-cli"
-version = "0.1.0"
+name = "coinhunter"
+version = "1.0.0"
 description = "CoinHunter trading CLI with user runtime data in ~/.coinhunter"
 readme = "README.md"
+license = {text = "MIT"}
 requires-python = ">=3.10"
 dependencies = [
  "ccxt>=4.4.0"
@@ -15,6 +16,9 @@ authors = [
  {name = "Tacit Lab", email = "ouyangcarlos@gmail.com"}
 ]

+[project.optional-dependencies]
+dev = ["pytest>=8.0", "ruff>=0.4.0", "mypy>=1.0"]
+
 [project.scripts]
 coinhunter = "coinhunter.cli:main"

@@ -23,3 +27,17 @@ package-dir = {"" = "src"}

 [tool.setuptools.packages.find]
 where = ["src"]
+
+[tool.ruff]
+line-length = 120
+target-version = "py310"
+
+[tool.ruff.lint]
+select = ["E", "F", "W", "I"]
+ignore = ["E501"]
+
+[tool.mypy]
+python_version = "3.10"
+warn_return_any = true
+warn_unused_configs = true
+ignore_missing_imports = true
--- a/scripts/install_local.sh
+++ b/scripts/install_local.sh
@@ -1,10 +1,10 @@
 #!/usr/bin/env bash
 set -euo pipefail

-APP_HOME="${COINHUNTER_APP_HOME:-$HOME/.local/share/coinhunter-cli}"
-VENV_DIR="$APP_HOME/venv"
+# Standard local install using pip editable mode.
+# This creates the 'coinhunter' console script entry point as defined in pyproject.toml.
+
 BIN_DIR="${COINHUNTER_BIN_DIR:-$HOME/.local/bin}"
-LAUNCHER="$BIN_DIR/coinhunter"
 PYTHON_BIN="${PYTHON:-}"

 if [[ -z "$PYTHON_BIN" ]]; then
@@ -18,18 +18,13 @@ if [[ -z "$PYTHON_BIN" ]]; then
  fi
 fi

-mkdir -p "$APP_HOME" "$BIN_DIR"
+mkdir -p "$BIN_DIR"

-"$PYTHON_BIN" -m venv "$VENV_DIR"
-"$VENV_DIR/bin/python" -m pip install --upgrade pip setuptools wheel
-"$VENV_DIR/bin/python" -m pip install --upgrade "$(pwd)"
+"$PYTHON_BIN" -m pip install --upgrade pip setuptools wheel
+"$PYTHON_BIN" -m pip install --upgrade -e "$(pwd)[dev]"

-cat >"$LAUNCHER" <<EOF
-#!/usr/bin/env bash
-exec "$VENV_DIR/bin/coinhunter" "\$@"
-EOF
-chmod +x "$LAUNCHER"
-
-echo "Installed coinhunter into:"
-echo "  venv: $VENV_DIR"
-echo "  launcher: $LAUNCHER"
+echo "Installed coinhunter in editable mode."
+echo "  python: $PYTHON_BIN"
+echo "  entrypoint: $(command -v coinhunter || echo 'not in PATH')"
+echo ""
+echo "Make sure '$BIN_DIR' is in your PATH if the entrypoint is not found."
--- a/src/coinhunter/main.py
+++ b/src/coinhunter/main.py
@@ -1,2 +1,3 @@
 from .cli import main
+
 raise SystemExit(main())
--- a/src/coinhunter/cli.py
+++ b/src/coinhunter/cli.py
@@ -16,8 +16,8 @@ MODULE_MAP = {
    "market-probe": "commands.market_probe",
    "paths": "commands.paths",
    "precheck": "commands.precheck",
-    "review-context": "review_context",
-    "review-engine": "review_engine",
+    "review-context": "commands.review_context",
+    "review-engine": "commands.review_engine",
    "rotate-external-gate-log": "commands.rotate_external_gate_log",
    "smart-executor": "commands.smart_executor",
 }
@@ -84,14 +84,17 @@ def build_parser() -> argparse.ArgumentParser:
            "  coinhunter exec overview\n"
            "  coinhunter exec hold\n"
            "  coinhunter exec --analysis '...' --reasoning '...' buy ENJUSDT 50\n"
+            "  coinhunter exec orders\n"
+            "  coinhunter exec order-status ENJUSDT 123456\n"
+            "  coinhunter exec cancel ENJUSDT 123456\n"
            "  coinhunter pre\n"
-            "  coinhunter pre --ack '分析完成：HOLD'\n"
+            "  coinhunter pre --ack 'Analysis complete: HOLD'\n"
            "  coinhunter gate\n"
            "  coinhunter review 12\n"
            "  coinhunter recap 12\n"
            "  coinhunter probe bybit-ticker BTCUSDT\n"
            "\n"
-            "Preferred exec verbs are bal, overview, hold, buy, flat, and rotate.\n"
+            "Preferred exec verbs are bal, overview, hold, buy, flat, rotate, orders, order-status, and cancel.\n"
            "Legacy command names remain supported for backward compatibility.\n"
        ),
    )
--- a/src/coinhunter/commands/check_api.py
+++ b/src/coinhunter/commands/check_api.py
@@ -1,7 +1,10 @@
 #!/usr/bin/env python3
-"""检查自动交易的环境配置是否就绪"""
+"""Check whether the trading environment is ready and API permissions are sufficient."""
+import json
 import os

+import ccxt
+
 from ..runtime import load_env_file


@@ -12,14 +15,41 @@ def main():
    secret = os.getenv("BINANCE_API_SECRET", "")

    if not api_key or api_key.startswith("***") or api_key.startswith("your_"):
-        print("❌ 未配置 BINANCE_API_KEY")
+        print(json.dumps({"ok": False, "error": "BINANCE_API_KEY not configured"}, ensure_ascii=False))
        return 1
    if not secret or secret.startswith("***") or secret.startswith("your_"):
-        print("❌ 未配置 BINANCE_API_SECRET")
+        print(json.dumps({"ok": False, "error": "BINANCE_API_SECRET not configured"}, ensure_ascii=False))
        return 1

-    print("✅ API 配置正常")
-    return 0
+    try:
+        ex = ccxt.binance({
+            "apiKey": api_key,
+            "secret": secret,
+            "options": {"defaultType": "spot"},
+            "enableRateLimit": True,
+        })
+        balance = ex.fetch_balance()
+    except Exception as e:
+        print(json.dumps({"ok": False, "error": f"Failed to connect or fetch balance: {e}"}, ensure_ascii=False))
+        return 1
+
+    read_permission = bool(balance and isinstance(balance, dict))
+
+    spot_trading_enabled = None
+    try:
+        restrictions = ex.sapi_get_account_api_restrictions()
+        spot_trading_enabled = restrictions.get("enableSpotAndMarginTrading") or restrictions.get("enableSpotTrading")
+    except Exception:
+        pass
+
+    report = {
+        "ok": read_permission,
+        "read_permission": read_permission,
+        "spot_trading_enabled": spot_trading_enabled,
+        "note": "spot_trading_enabled may be null if the key lacks permission to query restrictions; it does not necessarily mean trading is disabled.",
+    }
+    print(json.dumps(report, ensure_ascii=False, indent=2))
+    return 0 if read_permission else 1


 if __name__ == "__main__":
--- a/src/coinhunter/commands/doctor.py
+++ b/src/coinhunter/commands/doctor.py
@@ -10,7 +10,6 @@ import sys

 from ..runtime import ensure_runtime_dirs, get_runtime_paths, load_env_file, resolve_hermes_executable

-
 REQUIRED_ENV_VARS = ["BINANCE_API_KEY", "BINANCE_API_SECRET"]


--- a/src/coinhunter/commands/external_gate.py
+++ b/src/coinhunter/commands/external_gate.py
@@ -5,13 +5,14 @@ import subprocess
 import sys
 from datetime import datetime, timezone

-from ..runtime import ensure_runtime_dirs, get_runtime_paths, resolve_hermes_executable
+from ..runtime import ensure_runtime_dirs, get_runtime_paths
+
+
+def _paths():
+    return get_runtime_paths()
+

-PATHS = get_runtime_paths()
-STATE_DIR = PATHS.state_dir
-LOCK_FILE = PATHS.external_gate_lock
 COINHUNTER_MODULE = [sys.executable, "-m", "coinhunter"]
-TRADE_JOB_ID = "4e6593fff158"


 def utc_now():
@@ -19,7 +20,7 @@ def utc_now():


 def log(message: str):
-    print(f"[{utc_now()}] {message}")
+    print(f"[{utc_now()}] {message}", file=sys.stderr)


 def run_cmd(args: list[str]) -> subprocess.CompletedProcess:
@@ -30,51 +31,129 @@ def parse_json_output(text: str) -> dict:
    text = (text or "").strip()
    if not text:
        return {}
-    return json.loads(text)
+    return json.loads(text)  # type: ignore[no-any-return]
+
+
+def _load_config() -> dict:
+    config_path = _paths().config_file
+    if not config_path.exists():
+        return {}
+    try:
+        return json.loads(config_path.read_text(encoding="utf-8"))  # type: ignore[no-any-return]
+    except Exception:
+        return {}
+
+
+def _resolve_trigger_command(paths) -> list[str] | None:
+    config = _load_config()
+    gate_config = config.get("external_gate", {})
+
+    if "trigger_command" not in gate_config:
+        return None
+
+    trigger = gate_config["trigger_command"]
+
+    if trigger is None:
+        return None
+
+    if isinstance(trigger, str):
+        return [trigger]
+
+    if isinstance(trigger, list):
+        if not trigger:
+            return None
+        return [str(item) for item in trigger]
+
+    log(f"warn: unexpected trigger_command type {type(trigger).__name__}; skipping trigger")
+    return None


 def main():
-    ensure_runtime_dirs(PATHS)
-    with open(LOCK_FILE, "w", encoding="utf-8") as lockf:
+    paths = _paths()
+    ensure_runtime_dirs(paths)
+    result = {"ok": False, "triggered": False, "reason": "", "logs": []}
+    lock_file = paths.external_gate_lock
+
+    def append_log(msg: str):
+        log(msg)
+        result["logs"].append(msg)
+
+    with open(lock_file, "w", encoding="utf-8") as lockf:
        try:
            fcntl.flock(lockf.fileno(), fcntl.LOCK_EX | fcntl.LOCK_NB)
        except BlockingIOError:
-            log("gate already running; skip")
+            append_log("gate already running; skip")
+            result["reason"] = "already_running"
+            print(json.dumps(result, ensure_ascii=False))
            return 0

        precheck = run_cmd(COINHUNTER_MODULE + ["precheck"])
        if precheck.returncode != 0:
-            log(f"precheck returned non-zero ({precheck.returncode}); stdout={precheck.stdout.strip()} stderr={precheck.stderr.strip()}")
+            append_log(f"precheck returned non-zero ({precheck.returncode}); stdout={precheck.stdout.strip()} stderr={precheck.stderr.strip()}")
+            result["reason"] = "precheck_failed"
+            print(json.dumps(result, ensure_ascii=False))
            return 1

        try:
            data = parse_json_output(precheck.stdout)
        except Exception as e:
-            log(f"failed to parse precheck JSON: {e}; raw={precheck.stdout.strip()[:1000]}")
+            append_log(f"failed to parse precheck JSON: {e}; raw={precheck.stdout.strip()[:1000]}")
+            result["reason"] = "precheck_parse_error"
+            print(json.dumps(result, ensure_ascii=False))
+            return 1
+
+        if not data.get("ok"):
+            append_log("precheck reported failure; skip model run")
+            result["reason"] = "precheck_not_ok"
+            print(json.dumps(result, ensure_ascii=False))
            return 1

        if not data.get("should_analyze"):
-            log("no trigger; skip model run")
+            append_log("no trigger; skip model run")
+            result["ok"] = True
+            result["reason"] = "no_trigger"
+            print(json.dumps(result, ensure_ascii=False))
            return 0

        if data.get("run_requested"):
-            log(f"trigger already queued at {data.get('run_requested_at')}; skip duplicate")
+            append_log(f"trigger already queued at {data.get('run_requested_at')}; skip duplicate")
+            result["ok"] = True
+            result["reason"] = "already_queued"
+            print(json.dumps(result, ensure_ascii=False))
            return 0

        mark = run_cmd(COINHUNTER_MODULE + ["precheck", "--mark-run-requested", "external-gate queued cron run"])
        if mark.returncode != 0:
-            log(f"failed to mark run requested; stdout={mark.stdout.strip()} stderr={mark.stderr.strip()}")
+            append_log(f"failed to mark run requested; stdout={mark.stdout.strip()} stderr={mark.stderr.strip()}")
+            result["reason"] = "mark_failed"
+            print(json.dumps(result, ensure_ascii=False))
            return 1

-        trigger = run_cmd([resolve_hermes_executable(PATHS), "cron", "run", TRADE_JOB_ID])
+        trigger_cmd = _resolve_trigger_command(paths)
+        if trigger_cmd is None:
+            append_log("trigger_command is disabled; skipping external trigger")
+            result["ok"] = True
+            result["reason"] = "trigger_disabled"
+            print(json.dumps(result, ensure_ascii=False))
+            return 0
+
+        trigger = run_cmd(trigger_cmd)
        if trigger.returncode != 0:
-            log(f"failed to trigger trade cron job; stdout={trigger.stdout.strip()} stderr={trigger.stderr.strip()}")
+            append_log(f"failed to trigger trade job; cmd={' '.join(trigger_cmd)}; stdout={trigger.stdout.strip()} stderr={trigger.stderr.strip()}")
+            result["reason"] = "trigger_failed"
+            print(json.dumps(result, ensure_ascii=False))
            return 1

        reasons = ", ".join(data.get("reasons", [])) or "unknown"
-        log(f"queued trade job {TRADE_JOB_ID}; reasons={reasons}")
+        append_log(f"queued trade job via {' '.join(trigger_cmd)}; reasons={reasons}")
        if trigger.stdout.strip():
-            log(trigger.stdout.strip())
+            append_log(trigger.stdout.strip())
+
+        result["ok"] = True
+        result["triggered"] = True
+        result["reason"] = reasons
+        result["command"] = trigger_cmd
+        print(json.dumps(result, ensure_ascii=False))
        return 0


--- a/src/coinhunter/commands/init_user_state.py
+++ b/src/coinhunter/commands/init_user_state.py
@@ -5,9 +5,9 @@ from pathlib import Path

 from ..runtime import ensure_runtime_dirs, get_runtime_paths

-PATHS = get_runtime_paths()
-ROOT = PATHS.root
-CACHE_DIR = PATHS.cache_dir
+
+def _paths():
+    return get_runtime_paths()


 def now_iso():
@@ -22,30 +22,60 @@ def ensure_file(path: Path, payload: dict):


 def main():
-    ensure_runtime_dirs(PATHS)
+    paths = _paths()
+    ensure_runtime_dirs(paths)

    created = []
    ts = now_iso()

    templates = {
-        ROOT / "config.json": {
+        paths.root / "config.json": {
            "default_exchange": "bybit",
            "default_quote_currency": "USDT",
            "timezone": "Asia/Shanghai",
            "preferred_chains": ["solana", "base"],
+            "external_gate": {
+                "trigger_command": None,
+                "_comment": "Set to a command list like ['hermes', 'cron', 'run', 'JOB_ID'] or null to disable"
+            },
+            "trading": {
+                "usdt_buffer_pct": 0.03,
+                "min_remaining_dust_usdt": 1.0,
+                "_comment": "Adjust buffer and dust thresholds for your account size"
+            },
+            "precheck": {
+                "base_price_move_trigger_pct": 0.025,
+                "base_pnl_trigger_pct": 0.03,
+                "base_portfolio_move_trigger_pct": 0.03,
+                "base_candidate_score_trigger_ratio": 1.15,
+                "base_force_analysis_after_minutes": 180,
+                "base_cooldown_minutes": 45,
+                "top_candidates": 10,
+                "min_actionable_usdt": 12.0,
+                "min_real_position_value_usdt": 8.0,
+                "blacklist": ["USDC", "BUSD", "TUSD", "FDUSD", "USTC", "PAXG"],
+                "hard_stop_pct": -0.08,
+                "hard_moon_pct": 0.25,
+                "min_change_pct": 1.0,
+                "max_price_cap": None,
+                "hard_reason_dedup_minutes": 15,
+                "max_pending_trigger_minutes": 30,
+                "max_run_request_minutes": 20,
+                "_comment": "Tune trigger sensitivity without redeploying code"
+            },
            "created_at": ts,
            "updated_at": ts,
        },
-        ROOT / "accounts.json": {
+        paths.root / "accounts.json": {
            "accounts": []
        },
-        ROOT / "positions.json": {
+        paths.root / "positions.json": {
            "positions": []
        },
-        ROOT / "watchlist.json": {
+        paths.root / "watchlist.json": {
            "watchlist": []
        },
-        ROOT / "notes.json": {
+        paths.root / "notes.json": {
            "notes": []
        },
    }
@@ -55,9 +85,9 @@ def main():
            created.append(str(path))

    print(json.dumps({
-        "root": str(ROOT),
+        "root": str(paths.root),
        "created": created,
-        "cache_dir": str(CACHE_DIR),
+        "cache_dir": str(paths.cache_dir),
    }, ensure_ascii=False, indent=2))


--- a/src/coinhunter/commands/market_probe.py
+++ b/src/coinhunter/commands/market_probe.py
@@ -2,7 +2,6 @@
 import argparse
 import json
 import os
-import sys
 import urllib.parse
 import urllib.request

--- a/src/coinhunter/commands/review_context.py
+++ b/src/coinhunter/commands/review_context.py
@@ -0,0 +1,34 @@
+#!/usr/bin/env python3
+"""CLI adapter for review context."""
+
+import json
+import sys
+
+from ..services.review_service import generate_review
+
+
+def main():
+    hours = int(sys.argv[1]) if len(sys.argv) > 1 else 12
+    review = generate_review(hours)
+    compact = {
+        "review_period_hours": review.get("review_period_hours", hours),
+        "review_timestamp": review.get("review_timestamp"),
+        "total_decisions": review.get("total_decisions", 0),
+        "total_trades": review.get("total_trades", 0),
+        "total_errors": review.get("total_errors", 0),
+        "stats": review.get("stats", {}),
+        "insights": review.get("insights", []),
+        "recommendations": review.get("recommendations", []),
+        "decision_quality_top": review.get("decision_quality", [])[:5],
+        "should_report": bool(
+            review.get("total_decisions", 0)
+            or review.get("total_trades", 0)
+            or review.get("total_errors", 0)
+            or review.get("insights")
+        ),
+    }
+    print(json.dumps(compact, ensure_ascii=False, indent=2))
+
+
+if __name__ == "__main__":
+    main()
--- a/src/coinhunter/commands/review_engine.py
+++ b/src/coinhunter/commands/review_engine.py
@@ -0,0 +1,24 @@
+#!/usr/bin/env python3
+"""CLI adapter for review engine."""
+
+import json
+import sys
+
+from ..services.review_service import generate_review, save_review
+
+
+def main():
+    try:
+        hours = int(sys.argv[1]) if len(sys.argv) > 1 else 1
+        review = generate_review(hours)
+        path = save_review(review)
+        print(json.dumps({"ok": True, "saved_path": path, "review": review}, ensure_ascii=False, indent=2))
+    except Exception as e:
+        from ..logger import log_error
+        log_error("review_engine", e)
+        print(json.dumps({"ok": False, "error": str(e)}, ensure_ascii=False), file=sys.stderr)
+        raise SystemExit(1)
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())
--- a/src/coinhunter/commands/rotate_external_gate_log.py
+++ b/src/coinhunter/commands/rotate_external_gate_log.py
@@ -1,27 +1,30 @@
 #!/usr/bin/env python3
 """Rotate external gate log using the user's logrotate config/state."""
+import json
 import shutil
 import subprocess

 from ..runtime import ensure_runtime_dirs, get_runtime_paths

-PATHS = get_runtime_paths()
-STATE_DIR = PATHS.state_dir
-LOGROTATE_STATUS = PATHS.logrotate_status
-LOGROTATE_CONF = PATHS.logrotate_config
-LOGS_DIR = PATHS.logs_dir
+
+def _paths():
+    return get_runtime_paths()


 def main():
-    ensure_runtime_dirs(PATHS)
+    paths = _paths()
+    ensure_runtime_dirs(paths)
    logrotate_bin = shutil.which("logrotate") or "/usr/sbin/logrotate"
-    cmd = [logrotate_bin, "-s", str(LOGROTATE_STATUS), str(LOGROTATE_CONF)]
+    cmd = [logrotate_bin, "-s", str(paths.logrotate_status), str(paths.logrotate_config)]
    result = subprocess.run(cmd, capture_output=True, text=True)
-    if result.stdout.strip():
-        print(result.stdout.strip())
-    if result.stderr.strip():
-        print(result.stderr.strip())
-    return result.returncode
+    output = {
+        "ok": result.returncode == 0,
+        "returncode": result.returncode,
+        "stdout": result.stdout.strip(),
+        "stderr": result.stderr.strip(),
+    }
+    print(json.dumps(output, ensure_ascii=False, indent=2))
+    return 0 if result.returncode == 0 else 1


 if __name__ == "__main__":
--- a/src/coinhunter/logger.py
+++ b/src/coinhunter/logger.py
@@ -2,28 +2,41 @@
 """Coin Hunter structured logger."""
 import json
 import traceback
-from datetime import datetime, timezone, timedelta

-from .runtime import get_runtime_paths
+__all__ = [
+    "SCHEMA_VERSION",
+    "log_decision",
+    "log_trade",
+    "log_snapshot",
+    "log_error",
+    "get_logs_by_date",
+    "get_logs_last_n_hours",
+]
+from datetime import datetime, timedelta, timezone

-LOG_DIR = get_runtime_paths().logs_dir
-SCHEMA_VERSION = 2
+from .runtime import get_runtime_paths, get_user_config
+
+SCHEMA_VERSION = get_user_config("logging.schema_version", 2)

 CST = timezone(timedelta(hours=8))


+def _log_dir():
+    return get_runtime_paths().logs_dir
+
+
 def bj_now():
    return datetime.now(CST)


 def ensure_dir():
-    LOG_DIR.mkdir(parents=True, exist_ok=True)
+    _log_dir().mkdir(parents=True, exist_ok=True)


 def _append_jsonl(prefix: str, payload: dict):
    ensure_dir()
    date_str = bj_now().strftime("%Y%m%d")
-    log_file = LOG_DIR / f"{prefix}_{date_str}.jsonl"
+    log_file = _log_dir() / f"{prefix}_{date_str}.jsonl"
    with open(log_file, "a", encoding="utf-8") as f:
        f.write(json.dumps(payload, ensure_ascii=False) + "\n")

@@ -42,8 +55,8 @@ def log_decision(data: dict):
    return log_event("decisions", data)


-def log_trade(action: str, symbol: str, qty: float = None, amount_usdt: float = None,
-              price: float = None, note: str = "", **extra):
+def log_trade(action: str, symbol: str, qty: float | None = None, amount_usdt: float | None = None,
+              price: float | None = None, note: str = "", **extra):
    payload = {
        "action": action,
        "symbol": symbol,
@@ -71,10 +84,10 @@ def log_error(where: str, error: Exception | str, **extra):
    return log_event("errors", payload)


-def get_logs_by_date(log_type: str, date_str: str = None) -> list:
+def get_logs_by_date(log_type: str, date_str: str | None = None) -> list:
    if date_str is None:
        date_str = bj_now().strftime("%Y%m%d")
-    log_file = LOG_DIR / f"{log_type}_{date_str}.jsonl"
+    log_file = _log_dir() / f"{log_type}_{date_str}.jsonl"
    if not log_file.exists():
        return []
    entries = []
--- a/src/coinhunter/precheck.py
+++ b/src/coinhunter/precheck.py
@@ -13,13 +13,6 @@ from importlib import import_module
 from .services.precheck_service import run as _run_service

 _CORE_EXPORTS = {
-    "PATHS",
-    "BASE_DIR",
-    "STATE_DIR",
-    "STATE_FILE",
-    "POSITIONS_FILE",
-    "CONFIG_FILE",
-    "ENV_FILE",
    "BASE_PRICE_MOVE_TRIGGER_PCT",
    "BASE_PNL_TRIGGER_PCT",
    "BASE_PORTFOLIO_MOVE_TRIGGER_PCT",
@@ -64,14 +57,28 @@ _CORE_EXPORTS = {
    "analyze_trigger",
    "update_state_after_observation",
 }
+
+# Path-related exports are now lazy in precheck_core
+_PATH_EXPORTS = {
+    "PATHS",
+    "BASE_DIR",
+    "STATE_DIR",
+    "STATE_FILE",
+    "POSITIONS_FILE",
+    "CONFIG_FILE",
+    "ENV_FILE",
+}
+
 _STATE_EXPORTS = {"mark_run_requested", "ack_analysis"}

-__all__ = sorted(_CORE_EXPORTS | _STATE_EXPORTS | {"main"})
+__all__ = sorted(_CORE_EXPORTS | _PATH_EXPORTS | _STATE_EXPORTS | {"main"})


 def __getattr__(name: str):
    if name in _CORE_EXPORTS:
        return getattr(import_module(".services.precheck_core", __package__), name)
+    if name in _PATH_EXPORTS:
+        return getattr(import_module(".services.precheck_core", __package__), name)
    if name in _STATE_EXPORTS:
        return getattr(import_module(".services.precheck_state", __package__), name)
    raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
--- a/src/coinhunter/review_context.py
+++ b/src/coinhunter/review_context.py
@@ -1,32 +1,12 @@
 #!/usr/bin/env python3
-import json
-import sys
+"""Backward-compatible facade for review context.

-from . import review_engine
+The executable implementation lives in ``coinhunter.commands.review_context``.
+"""

+from __future__ import annotations

-def main():
-    hours = int(sys.argv[1]) if len(sys.argv) > 1 else 12
-    review = review_engine.generate_review(hours)
-    compact = {
-        "review_period_hours": review.get("review_period_hours", hours),
-        "review_timestamp": review.get("review_timestamp"),
-        "total_decisions": review.get("total_decisions", 0),
-        "total_trades": review.get("total_trades", 0),
-        "total_errors": review.get("total_errors", 0),
-        "stats": review.get("stats", {}),
-        "insights": review.get("insights", []),
-        "recommendations": review.get("recommendations", []),
-        "decision_quality_top": review.get("decision_quality", [])[:5],
-        "should_report": bool(
-            review.get("total_decisions", 0)
-            or review.get("total_trades", 0)
-            or review.get("total_errors", 0)
-            or review.get("insights")
-        ),
-    }
-    print(json.dumps(compact, ensure_ascii=False, indent=2))
-
+from .commands.review_context import main

 if __name__ == "__main__":
    main()
--- a/src/coinhunter/review_engine.py
+++ b/src/coinhunter/review_engine.py
@@ -1,312 +1,48 @@
 #!/usr/bin/env python3
-"""Coin Hunter hourly review engine."""
-import json
-import os
-import sys
-from datetime import datetime, timezone, timedelta
-from pathlib import Path
+"""Backward-compatible facade for review engine.

-import ccxt
+The executable implementation lives in ``coinhunter.commands.review_engine``.
+Core logic is in ``coinhunter.services.review_service``.
+"""

-from .logger import get_logs_last_n_hours, log_error
-from .runtime import get_runtime_paths, load_env_file
+from __future__ import annotations

-PATHS = get_runtime_paths()
-ENV_FILE = PATHS.env_file
-REVIEW_DIR = PATHS.reviews_dir
+from importlib import import_module

-CST = timezone(timedelta(hours=8))
+# Re-export service functions for backward compatibility
+_EXPORT_MAP = {
+    "load_env": (".services.review_service", "load_env"),
+    "get_exchange": (".services.review_service", "get_exchange"),
+    "ensure_review_dir": (".services.review_service", "ensure_review_dir"),
+    "norm_symbol": (".services.review_service", "norm_symbol"),
+    "fetch_current_price": (".services.review_service", "fetch_current_price"),
+    "analyze_trade": (".services.review_service", "analyze_trade"),
+    "analyze_hold_passes": (".services.review_service", "analyze_hold_passes"),
+    "analyze_cash_misses": (".services.review_service", "analyze_cash_misses"),
+    "generate_review": (".services.review_service", "generate_review"),
+    "save_review": (".services.review_service", "save_review"),
+    "print_review": (".services.review_service", "print_review"),
+}
+
+__all__ = sorted(set(_EXPORT_MAP) | {"main"})


-def load_env():
-    load_env_file(PATHS)
+def __getattr__(name: str):
+    if name not in _EXPORT_MAP:
+        raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
+    module_name, attr_name = _EXPORT_MAP[name]
+    module = import_module(module_name, __package__)
+    return getattr(module, attr_name)


-def get_exchange():
-    load_env()
-    ex = ccxt.binance({
-        "apiKey": os.getenv("BINANCE_API_KEY"),
-        "secret": os.getenv("BINANCE_API_SECRET"),
-        "options": {"defaultType": "spot"},
-        "enableRateLimit": True,
-    })
-    ex.load_markets()
-    return ex
-
-
-def ensure_review_dir():
-    REVIEW_DIR.mkdir(parents=True, exist_ok=True)
-
-
-def norm_symbol(symbol: str) -> str:
-    s = symbol.upper().replace("-", "").replace("_", "")
-    if "/" in s:
-        return s
-    if s.endswith("USDT"):
-        return s[:-4] + "/USDT"
-    return s
-
-
-def fetch_current_price(ex, symbol: str):
-    try:
-        return float(ex.fetch_ticker(norm_symbol(symbol))["last"])
-    except Exception:
-        return None
-
-
-def analyze_trade(trade: dict, ex) -> dict:
-    symbol = trade.get("symbol")
-    price = trade.get("price")
-    action = trade.get("action", "")
-    current_price = fetch_current_price(ex, symbol) if symbol else None
-    pnl_estimate = None
-    outcome = "neutral"
-    if price and current_price and symbol:
-        change_pct = (current_price - float(price)) / float(price) * 100
-        if action == "BUY":
-            pnl_estimate = round(change_pct, 2)
-            outcome = "good" if change_pct > 2 else "bad" if change_pct < -2 else "neutral"
-        elif action == "SELL_ALL":
-            pnl_estimate = round(-change_pct, 2)
-            # Lowered missed threshold: >2% is a missed opportunity in short-term trading
-            outcome = "good" if change_pct < -2 else "missed" if change_pct > 2 else "neutral"
-    return {
-        "timestamp": trade.get("timestamp"),
-        "symbol": symbol,
-        "action": action,
-        "decision_id": trade.get("decision_id"),
-        "execution_price": price,
-        "current_price": current_price,
-        "pnl_estimate_pct": pnl_estimate,
-        "outcome_assessment": outcome,
-    }
-
-
-def analyze_hold_passes(decisions: list, ex) -> list:
-    """Check HOLD decisions where an opportunity was explicitly PASSed but later rallied."""
-    misses = []
-    for d in decisions:
-        if d.get("decision") != "HOLD":
-            continue
-        analysis = d.get("analysis")
-        if not isinstance(analysis, dict):
-            continue
-        opportunities = analysis.get("opportunities_evaluated", [])
-        market_snapshot = d.get("market_snapshot", {})
-        if not opportunities or not market_snapshot:
-            continue
-        for op in opportunities:
-            verdict = op.get("verdict", "")
-            if "PASS" not in verdict and "pass" not in verdict:
-                continue
-            symbol = op.get("symbol", "")
-            # Try to extract decision-time price from market_snapshot
-            snap = market_snapshot.get(symbol) or market_snapshot.get(symbol.replace("/", ""))
-            if not snap:
-                continue
-            decision_price = None
-            if isinstance(snap, dict):
-                decision_price = float(snap.get("lastPrice", 0)) or float(snap.get("last", 0))
-            elif isinstance(snap, (int, float, str)):
-                decision_price = float(snap)
-            if not decision_price:
-                continue
-            current_price = fetch_current_price(ex, symbol)
-            if not current_price:
-                continue
-            change_pct = (current_price - decision_price) / decision_price * 100
-            if change_pct > 3:  # >3% rally after being passed = missed watch
-                misses.append({
-                    "timestamp": d.get("timestamp"),
-                    "symbol": symbol,
-                    "decision_price": round(decision_price, 8),
-                    "current_price": round(current_price, 8),
-                    "change_pct": round(change_pct, 2),
-                    "verdict_snippet": verdict[:80],
-                })
-    return misses
-
-
-def analyze_cash_misses(decisions: list, ex) -> list:
-    """If portfolio was mostly USDT but a watchlist coin rallied >5%, flag it."""
-    misses = []
-    watchlist = set()
-    for d in decisions:
-        snap = d.get("market_snapshot", {})
-        if isinstance(snap, dict):
-            for k in snap.keys():
-                if k.endswith("USDT"):
-                    watchlist.add(k)
-    for d in decisions:
-        ts = d.get("timestamp")
-        balances = d.get("balances") or d.get("balances_before", {})
-        if not balances:
-            continue
-        total = sum(float(v) if isinstance(v, (int, float, str)) else 0 for v in balances.values())
-        usdt = float(balances.get("USDT", 0))
-        if total == 0 or (usdt / total) < 0.9:
-            continue
-        # Portfolio mostly cash — check watchlist performance
-        snap = d.get("market_snapshot", {})
-        if not isinstance(snap, dict):
-            continue
-        for symbol, data in snap.items():
-            if not symbol.endswith("USDT"):
-                continue
-            decision_price = None
-            if isinstance(data, dict):
-                decision_price = float(data.get("lastPrice", 0)) or float(data.get("last", 0))
-            elif isinstance(data, (int, float, str)):
-                decision_price = float(data)
-            if not decision_price:
-                continue
-            current_price = fetch_current_price(ex, symbol)
-            if not current_price:
-                continue
-            change_pct = (current_price - decision_price) / decision_price * 100
-            if change_pct > 5:
-                misses.append({
-                    "timestamp": ts,
-                    "symbol": symbol,
-                    "decision_price": round(decision_price, 8),
-                    "current_price": round(current_price, 8),
-                    "change_pct": round(change_pct, 2),
-                })
-    # Deduplicate by symbol keeping the worst miss
-    seen = {}
-    for m in misses:
-        sym = m["symbol"]
-        if sym not in seen or m["change_pct"] > seen[sym]["change_pct"]:
-            seen[sym] = m
-    return list(seen.values())
-
-
-def generate_review(hours: int = 1) -> dict:
-    decisions = get_logs_last_n_hours("decisions", hours)
-    trades = get_logs_last_n_hours("trades", hours)
-    errors = get_logs_last_n_hours("errors", hours)
-
-    review = {
-        "review_period_hours": hours,
-        "review_timestamp": datetime.now(CST).isoformat(),
-        "total_decisions": len(decisions),
-        "total_trades": len(trades),
-        "total_errors": len(errors),
-        "decision_quality": [],
-        "stats": {},
-        "insights": [],
-        "recommendations": [],
-    }
-
-    if not decisions and not trades:
-        review["insights"].append("本周期无决策/交易记录")
-        return review
-
-    ex = get_exchange()
-    outcomes = {"good": 0, "neutral": 0, "bad": 0, "missed": 0}
-    pnl_samples = []
-
-    for trade in trades:
-        analysis = analyze_trade(trade, ex)
-        review["decision_quality"].append(analysis)
-        outcomes[analysis["outcome_assessment"]] += 1
-        if analysis["pnl_estimate_pct"] is not None:
-            pnl_samples.append(analysis["pnl_estimate_pct"])
-
-    # New: analyze missed opportunities from HOLD / cash decisions
-    hold_pass_misses = analyze_hold_passes(decisions, ex)
-    cash_misses = analyze_cash_misses(decisions, ex)
-    total_missed = outcomes["missed"] + len(hold_pass_misses) + len(cash_misses)
-
-    review["stats"] = {
-        "good_decisions": outcomes["good"],
-        "neutral_decisions": outcomes["neutral"],
-        "bad_decisions": outcomes["bad"],
-        "missed_opportunities": total_missed,
-        "missed_sell_all": outcomes["missed"],
-        "missed_hold_passes": len(hold_pass_misses),
-        "missed_cash_sits": len(cash_misses),
-        "avg_estimated_edge_pct": round(sum(pnl_samples) / len(pnl_samples), 2) if pnl_samples else None,
-    }
-
-    if errors:
-        review["insights"].append(f"本周期出现 {len(errors)} 次执行/系统错误，健壮性需优先关注")
-    if outcomes["bad"] > outcomes["good"]:
-        review["insights"].append("最近交易质量偏弱，建议降低交易频率或提高入场门槛")
-    if total_missed > 0:
-        parts = []
-        if outcomes["missed"]:
-            parts.append(f"卖出后继续上涨 {outcomes['missed']} 次")
-        if hold_pass_misses:
-            parts.append(f"PASS 后错失 {len(hold_pass_misses)} 次")
-        if cash_misses:
-            parts.append(f"空仓观望错失 {len(cash_misses)} 次")
-        review["insights"].append("存在错失机会: " + "，".join(parts) + "，建议放宽趋势跟随或入场条件")
-    if outcomes["good"] >= max(1, outcomes["bad"] + total_missed):
-        review["insights"].append("近期决策总体可接受")
-    if not trades and decisions:
-        review["insights"].append("有决策无成交，可能是观望、最小成交额限制或执行被拦截")
-    if len(trades) < len(decisions) * 0.1 and decisions:
-        review["insights"].append("大量决策未转化为交易，需检查执行门槛（最小成交额/精度/手续费缓冲）是否过高")
-    if hold_pass_misses:
-        for m in hold_pass_misses[:3]:
-            review["insights"].append(f"HOLD 时 PASS 了 {m['symbol']}，之后上涨 {m['change_pct']}%")
-    if cash_misses:
-        for m in cash_misses[:3]:
-            review["insights"].append(f"持仓以 USDT 为主时 {m['symbol']} 上涨 {m['change_pct']}%")
-
-    review["recommendations"] = [
-        "优先检查最小成交额/精度拒单是否影响小资金执行",
-        "若连续两个复盘周期 edge 为负，下一小时减少换仓频率",
-        "若错误日志增加，优先进入防守模式（多持 USDT）",
-    ]
-    return review
-
-
-def save_review(review: dict):
-    ensure_review_dir()
-    ts = datetime.now(CST).strftime("%Y%m%d_%H%M%S")
-    path = REVIEW_DIR / f"review_{ts}.json"
-    path.write_text(json.dumps(review, indent=2, ensure_ascii=False), encoding="utf-8")
-    return str(path)
-
-
-def print_review(review: dict):
-    print("=" * 50)
-    print("📊 Coin Hunter 小时复盘报告")
-    print(f"复盘时间: {review['review_timestamp']}")
-    print(f"统计周期: 过去 {review['review_period_hours']} 小时")
-    print(f"总决策数: {review['total_decisions']} | 总交易数: {review['total_trades']} | 总错误数: {review['total_errors']}")
-    stats = review.get("stats", {})
-    print("\n决策质量统计:")
-    print(f"  ✓ 优秀: {stats.get('good_decisions', 0)}")
-    print(f"  ○ 中性: {stats.get('neutral_decisions', 0)}")
-    print(f"  ✗ 失误: {stats.get('bad_decisions', 0)}")
-    print(f"  ↗ 错过机会: {stats.get('missed_opportunities', 0)}")
-    if stats.get("avg_estimated_edge_pct") is not None:
-        print(f"  平均估计 edge: {stats['avg_estimated_edge_pct']}%")
-    if review.get("insights"):
-        print("\n💡 见解:")
-        for item in review["insights"]:
-            print(f"  • {item}")
-    if review.get("recommendations"):
-        print("\n🔧 优化建议:")
-        for item in review["recommendations"]:
-            print(f"  • {item}")
-    print("=" * 50)
+def __dir__():
+    return sorted(set(globals()) | set(__all__))


 def main():
-    try:
-        hours = int(sys.argv[1]) if len(sys.argv) > 1 else 1
-        review = generate_review(hours)
-        path = save_review(review)
-        print_review(review)
-        print(f"复盘已保存至: {path}")
-    except Exception as e:
-        log_error("review_engine", e)
-        raise
+    from .commands.review_engine import main as _main
+    return _main()


 if __name__ == "__main__":
-    main()
+    raise SystemExit(main())
--- a/src/coinhunter/runtime.py
+++ b/src/coinhunter/runtime.py
@@ -2,6 +2,7 @@

 from __future__ import annotations

+import json
 import os
 import shutil
 from dataclasses import asdict, dataclass
@@ -24,6 +25,7 @@ class RuntimePaths:
    positions_lock: Path
    executions_lock: Path
    precheck_state_file: Path
+    precheck_state_lock: Path
    external_gate_lock: Path
    logrotate_config: Path
    logrotate_status: Path
@@ -64,6 +66,7 @@ def get_runtime_paths() -> RuntimePaths:
        positions_lock=root / "positions.lock",
        executions_lock=root / "executions.lock",
        precheck_state_file=state_dir / "precheck_state.json",
+        precheck_state_lock=state_dir / "precheck_state.lock",
        external_gate_lock=state_dir / "external_gate.lock",
        logrotate_config=root / "logrotate_external_gate.conf",
        logrotate_status=state_dir / "logrotate_external_gate.status",
@@ -105,3 +108,20 @@ def mask_secret(value: str | None, *, tail: int = 4) -> str:
    if len(value) <= tail:
        return "*" * len(value)
    return "*" * max(4, len(value) - tail) + value[-tail:]
+
+
+def get_user_config(key: str, default=None):
+    """Read a dotted key from the user config file."""
+    paths = get_runtime_paths()
+    try:
+        config = json.loads(paths.config_file.read_text(encoding="utf-8"))
+    except Exception:
+        return default
+    for part in key.split("."):
+        if isinstance(config, dict):
+            config = config.get(part)
+            if config is None:
+                return default
+        else:
+            return default
+    return config if config is not None else default
--- a/src/coinhunter/services/adaptive_profile.py
+++ b/src/coinhunter/services/adaptive_profile.py
@@ -31,7 +31,7 @@ def build_adaptive_profile(snapshot: dict):
    dust_mode = free_usdt < MIN_ACTIONABLE_USDT and largest_position_value < MIN_REAL_POSITION_VALUE_USDT

    price_trigger = BASE_PRICE_MOVE_TRIGGER_PCT
-    pnl_trigger = BASE_PN_L_TRIGGER_PCT
+    pnl_trigger = BASE_PNL_TRIGGER_PCT
    portfolio_trigger = BASE_PORTFOLIO_MOVE_TRIGGER_PCT
    candidate_ratio = BASE_CANDIDATE_SCORE_TRIGGER_RATIO
    force_minutes = BASE_FORCE_ANALYSIS_AFTER_MINUTES
--- a/src/coinhunter/services/candidate_scoring.py
+++ b/src/coinhunter/services/candidate_scoring.py
@@ -63,7 +63,7 @@ def top_candidates_from_tickers(tickers: dict):
        })
    candidates.sort(key=lambda x: x["score"], reverse=True)
    global_top = candidates[:TOP_CANDIDATES]
-    layers = {"major": [], "mid": [], "meme": []}
+    layers: dict[str, list[dict]] = {"major": [], "mid": [], "meme": []}
    for c in candidates:
        layers[c["band"]].append(c)
    for k in layers:
--- a/src/coinhunter/services/exchange_service.py
+++ b/src/coinhunter/services/exchange_service.py
@@ -1,25 +1,47 @@
 """Exchange helpers (ccxt, markets, balances, order prep)."""
 import math
 import os
+import time
+
+__all__ = [
+    "load_env",
+    "get_exchange",
+    "norm_symbol",
+    "storage_symbol",
+    "fetch_balances",
+    "build_market_snapshot",
+    "market_and_ticker",
+    "floor_to_step",
+    "prepare_buy_quantity",
+    "prepare_sell_quantity",
+]

 import ccxt

-from ..runtime import get_runtime_paths, load_env_file
-from .trade_common import log
+from ..runtime import get_runtime_paths, get_user_config, load_env_file

-PATHS = get_runtime_paths()
+_exchange_cache = None
+_exchange_cached_at = None
+
+CACHE_TTL_SECONDS = 3600


 def load_env():
-    load_env_file(PATHS)
+    load_env_file(get_runtime_paths())


-def get_exchange():
+def get_exchange(force_new: bool = False):
+    global _exchange_cache, _exchange_cached_at
+    now = time.time()
+    if not force_new and _exchange_cache is not None and _exchange_cached_at is not None:
+        ttl = get_user_config("exchange.cache_ttl_seconds", CACHE_TTL_SECONDS)
+        if now - _exchange_cached_at < ttl:
+            return _exchange_cache
    load_env()
    api_key = os.getenv("BINANCE_API_KEY")
    secret = os.getenv("BINANCE_API_SECRET")
    if not api_key or not secret:
-        raise RuntimeError("缺少 BINANCE_API_KEY 或 BINANCE_API_SECRET")
+        raise RuntimeError("Missing BINANCE_API_KEY or BINANCE_API_SECRET")
    ex = ccxt.binance(
        {
            "apiKey": api_key,
@@ -29,6 +51,8 @@ def get_exchange():
        }
    )
    ex.load_markets()
+    _exchange_cache = ex
+    _exchange_cached_at = now
    return ex


@@ -38,7 +62,7 @@ def norm_symbol(symbol: str) -> str:
        return s
    if s.endswith("USDT"):
        return s[:-4] + "/USDT"
-    raise ValueError(f"不支持的 symbol: {symbol}")
+    raise ValueError(f"Unsupported symbol: {symbol}")


 def storage_symbol(symbol: str) -> str:
@@ -63,7 +87,8 @@ def build_market_snapshot(ex):
        if price is None or float(price) <= 0:
            continue
        vol = float(t.get("quoteVolume") or 0)
-        if vol < 200_000:
+        min_volume = get_user_config("exchange.min_quote_volume", 200_000)
+        if vol < min_volume:
            continue
        base = sym.replace("/", "")
        snapshot[base] = {
@@ -95,17 +120,17 @@ def prepare_buy_quantity(ex, symbol: str, amount_usdt: float):
    sym, market, ticker = market_and_ticker(ex, symbol)
    ask = float(ticker.get("ask") or ticker.get("last") or 0)
    if ask <= 0:
-        raise RuntimeError(f"{sym} 无法获取有效 ask 价格")
+        raise RuntimeError(f"No valid ask price for {sym}")
    budget = amount_usdt * (1 - USDT_BUFFER_PCT)
    raw_qty = budget / ask
    qty = float(ex.amount_to_precision(sym, raw_qty))
    min_amt = (market.get("limits", {}).get("amount", {}) or {}).get("min") or 0
    min_cost = (market.get("limits", {}).get("cost", {}) or {}).get("min") or 0
    if min_amt and qty < float(min_amt):
-        raise RuntimeError(f"{sym} 买入数量 {qty} 小于最小数量 {min_amt}")
+        raise RuntimeError(f"Buy quantity {qty} for {sym} below minimum {min_amt}")
    est_cost = qty * ask
    if min_cost and est_cost < float(min_cost):
-        raise RuntimeError(f"{sym} 买入金额 ${est_cost:.4f} 小于最小成交额 ${float(min_cost):.4f}")
+        raise RuntimeError(f"Buy cost ${est_cost:.4f} for {sym} below minimum ${float(min_cost):.4f}")
    return sym, qty, ask, est_cost


@@ -113,13 +138,13 @@ def prepare_sell_quantity(ex, symbol: str, free_qty: float):
    sym, market, ticker = market_and_ticker(ex, symbol)
    bid = float(ticker.get("bid") or ticker.get("last") or 0)
    if bid <= 0:
-        raise RuntimeError(f"{sym} 无法获取有效 bid 价格")
+        raise RuntimeError(f"No valid bid price for {sym}")
    qty = float(ex.amount_to_precision(sym, free_qty))
    min_amt = (market.get("limits", {}).get("amount", {}) or {}).get("min") or 0
    min_cost = (market.get("limits", {}).get("cost", {}) or {}).get("min") or 0
    if min_amt and qty < float(min_amt):
-        raise RuntimeError(f"{sym} 卖出数量 {qty} 小于最小数量 {min_amt}")
+        raise RuntimeError(f"Sell quantity {qty} for {sym} below minimum {min_amt}")
    est_cost = qty * bid
    if min_cost and est_cost < float(min_cost):
-        raise RuntimeError(f"{sym} 卖出金额 ${est_cost:.4f} 小于最小成交额 ${float(min_cost):.4f}")
+        raise RuntimeError(f"Sell cost ${est_cost:.4f} for {sym} below minimum ${float(min_cost):.4f}")
    return sym, qty, bid, est_cost
--- a/src/coinhunter/services/execution_state.py
+++ b/src/coinhunter/services/execution_state.py
@@ -1,17 +1,25 @@
 """Execution state helpers (decision deduplication, executions.json)."""
 import hashlib

-from ..runtime import get_runtime_paths
-from .file_utils import load_json_locked, save_json_locked
-from .trade_common import bj_now_iso
+__all__ = [
+    "default_decision_id",
+    "load_executions",
+    "save_executions",
+    "record_execution_state",
+    "get_execution_state",
+]

-PATHS = get_runtime_paths()
-EXECUTIONS_FILE = PATHS.executions_file
-EXECUTIONS_LOCK = PATHS.executions_lock
+from ..runtime import get_runtime_paths
+from .file_utils import load_json_locked, read_modify_write_json, save_json_locked
+
+
+def _paths():
+    return get_runtime_paths()


 def default_decision_id(action: str, argv_tail: list[str]) -> str:
    from datetime import datetime
+
    from .trade_common import CST

    now = datetime.now(CST)
@@ -22,17 +30,24 @@ def default_decision_id(action: str, argv_tail: list[str]) -> str:


 def load_executions() -> dict:
-    return load_json_locked(EXECUTIONS_FILE, EXECUTIONS_LOCK, {"executions": {}}).get("executions", {})
+    paths = _paths()
+    data = load_json_locked(paths.executions_file, paths.executions_lock, {"executions": {}})
+    return data.get("executions", {})  # type: ignore[no-any-return]


 def save_executions(executions: dict):
-    save_json_locked(EXECUTIONS_FILE, EXECUTIONS_LOCK, {"executions": executions})
+    paths = _paths()
+    save_json_locked(paths.executions_file, paths.executions_lock, {"executions": executions})


 def record_execution_state(decision_id: str, payload: dict):
-    executions = load_executions()
-    executions[decision_id] = payload
-    save_executions(executions)
+    paths = _paths()
+    read_modify_write_json(
+        paths.executions_file,
+        paths.executions_lock,
+        {"executions": {}},
+        lambda data: data.setdefault("executions", {}).__setitem__(decision_id, payload) or data,
+    )


 def get_execution_state(decision_id: str):
--- a/src/coinhunter/services/file_utils.py
+++ b/src/coinhunter/services/file_utils.py
@@ -2,6 +2,14 @@
 import fcntl
 import json
 import os
+
+__all__ = [
+    "locked_file",
+    "atomic_write_json",
+    "load_json_locked",
+    "save_json_locked",
+    "read_modify_write_json",
+]
 from contextlib import contextmanager
 from pathlib import Path

@@ -9,13 +17,22 @@ from pathlib import Path
@contextmanager
 def locked_file(path: Path):
    path.parent.mkdir(parents=True, exist_ok=True)
-    with open(path, "a+", encoding="utf-8") as f:
-        fcntl.flock(f.fileno(), fcntl.LOCK_EX)
-        f.seek(0)
-        yield f
-        f.flush()
-        os.fsync(f.fileno())
-        fcntl.flock(f.fileno(), fcntl.LOCK_UN)
+    fd = None
+    try:
+        fd = os.open(path, os.O_RDWR | os.O_CREAT)
+        fcntl.flock(fd, fcntl.LOCK_EX)
+        yield fd
+    finally:
+        if fd is not None:
+            try:
+                os.fsync(fd)
+            except Exception:
+                pass
+            try:
+                fcntl.flock(fd, fcntl.LOCK_UN)
+            except Exception:
+                pass
+            os.close(fd)


 def atomic_write_json(path: Path, data: dict):
@@ -38,3 +55,22 @@ def load_json_locked(path: Path, lock_path: Path, default):
 def save_json_locked(path: Path, lock_path: Path, data: dict):
    with locked_file(lock_path):
        atomic_write_json(path, data)
+
+
+def read_modify_write_json(path: Path, lock_path: Path, default, modifier):
+    """Atomic read-modify-write under a single file lock.
+
+    Loads JSON from *path* (or uses *default* if missing/invalid),
+    calls ``modifier(data)``, then atomically writes the result back.
+    If *modifier* returns None, the mutated *data* is written.
+    """
+    with locked_file(lock_path):
+        if path.exists():
+            try:
+                data = json.loads(path.read_text(encoding="utf-8"))
+            except Exception:
+                data = default
+        else:
+            data = default
+        result = modifier(data)
+        atomic_write_json(path, result if result is not None else data)
--- a/src/coinhunter/services/market_data.py
+++ b/src/coinhunter/services/market_data.py
@@ -7,15 +7,12 @@ import os
 import ccxt

 from .data_utils import norm_symbol, to_float
-from .precheck_constants import BLACKLIST, MAX_PRICE_CAP, MIN_CHANGE_PCT
-from .time_utils import utc_now


 def get_exchange():
    from ..runtime import load_env_file
-    from .precheck_constants import ENV_FILE

-    load_env_file(ENV_FILE)
+    load_env_file()
    api_key = os.getenv("BINANCE_API_KEY")
    secret = os.getenv("BINANCE_API_SECRET")
    if not api_key or not secret:
--- a/src/coinhunter/services/portfolio_service.py
+++ b/src/coinhunter/services/portfolio_service.py
@@ -1,19 +1,47 @@
 """Portfolio state helpers (positions.json, reconcile with exchange)."""
 from ..runtime import get_runtime_paths
-from .file_utils import load_json_locked, save_json_locked
+
+__all__ = [
+    "load_positions",
+    "save_positions",
+    "update_positions",
+    "upsert_position",
+    "reconcile_positions_with_exchange",
+]
+from .file_utils import load_json_locked, read_modify_write_json, save_json_locked
 from .trade_common import bj_now_iso

-PATHS = get_runtime_paths()
-POSITIONS_FILE = PATHS.positions_file
-POSITIONS_LOCK = PATHS.positions_lock
+
+def _paths():
+    return get_runtime_paths()


 def load_positions() -> list:
-    return load_json_locked(POSITIONS_FILE, POSITIONS_LOCK, {"positions": []}).get("positions", [])
+    paths = _paths()
+    data = load_json_locked(paths.positions_file, paths.positions_lock, {"positions": []})
+    return data.get("positions", [])  # type: ignore[no-any-return]


 def save_positions(positions: list):
-    save_json_locked(POSITIONS_FILE, POSITIONS_LOCK, {"positions": positions})
+    paths = _paths()
+    save_json_locked(paths.positions_file, paths.positions_lock, {"positions": positions})
+
+
+def update_positions(modifier):
+    """Atomic read-modify-write for positions under a single lock.
+
+    *modifier* receives the current positions list and may mutate it in-place
+    or return a new list. If it returns None, the mutated list is saved.
+    """
+    paths = _paths()
+
+    def _modify(data):
+        positions = data.get("positions", [])
+        result = modifier(positions)
+        data["positions"] = result if result is not None else positions
+        return data
+
+    read_modify_write_json(paths.positions_file, paths.positions_lock, {"positions": []}, _modify)


 def upsert_position(positions: list, position: dict):
@@ -26,32 +54,43 @@ def upsert_position(positions: list, position: dict):
    return positions


-def reconcile_positions_with_exchange(ex, positions: list):
+def reconcile_positions_with_exchange(ex, positions_hint: list | None = None):
    from .exchange_service import fetch_balances

    balances = fetch_balances(ex)
-    existing_by_symbol = {p.get("symbol"): p for p in positions}
    reconciled = []
-    for asset, qty in balances.items():
-        if asset == "USDT":
-            continue
-        if qty <= 0:
-            continue
-        sym = f"{asset}USDT"
-        old = existing_by_symbol.get(sym, {})
-        reconciled.append(
-            {
-                "account_id": old.get("account_id", "binance-main"),
-                "symbol": sym,
-                "base_asset": asset,
-                "quote_asset": "USDT",
-                "market_type": "spot",
-                "quantity": qty,
-                "avg_cost": old.get("avg_cost"),
-                "opened_at": old.get("opened_at", bj_now_iso()),
-                "updated_at": bj_now_iso(),
-                "note": old.get("note", "Reconciled from Binance balances"),
-            }
-        )
-    save_positions(reconciled)
+
+    def _modify(data):
+        nonlocal reconciled
+        existing = data.get("positions", [])
+        existing_by_symbol = {p.get("symbol"): p for p in existing}
+        if positions_hint is not None:
+            existing_by_symbol.update({p.get("symbol"): p for p in positions_hint})
+        reconciled = []
+        for asset, qty in balances.items():
+            if asset == "USDT":
+                continue
+            if qty <= 0:
+                continue
+            sym = f"{asset}USDT"
+            old = existing_by_symbol.get(sym, {})
+            reconciled.append(
+                {
+                    "account_id": old.get("account_id", "binance-main"),
+                    "symbol": sym,
+                    "base_asset": asset,
+                    "quote_asset": "USDT",
+                    "market_type": "spot",
+                    "quantity": qty,
+                    "avg_cost": old.get("avg_cost"),
+                    "opened_at": old.get("opened_at", bj_now_iso()),
+                    "updated_at": bj_now_iso(),
+                    "note": old.get("note", "Reconciled from Binance balances"),
+                }
+            )
+        data["positions"] = reconciled
+        return data
+
+    paths = _paths()
+    read_modify_write_json(paths.positions_file, paths.positions_lock, {"positions": []}, _modify)
    return reconciled, balances
--- a/src/coinhunter/services/precheck_analysis.py
+++ b/src/coinhunter/services/precheck_analysis.py
@@ -3,7 +3,6 @@
 from __future__ import annotations

 from .time_utils import utc_iso
-from .trigger_analyzer import analyze_trigger


 def build_failure_payload(exc: Exception) -> dict:
@@ -18,5 +17,5 @@ def build_failure_payload(exc: Exception) -> dict:
        "soft_reasons": [],
        "soft_score": 0,
        "details": [str(exc)],
-        "compact_summary": f"预检查失败，转入深度分析兑底: {exc}",
+        "compact_summary": f"Precheck failed, falling back to deep analysis: {exc}",
    }
--- a/src/coinhunter/services/precheck_constants.py
+++ b/src/coinhunter/services/precheck_constants.py
@@ -1,31 +1,25 @@
-"""Precheck constants and runtime paths."""
+"""Precheck constants and thresholds."""

 from __future__ import annotations

-from ..runtime import get_runtime_paths
+from ..runtime import get_user_config

-PATHS = get_runtime_paths()
-BASE_DIR = PATHS.root
-STATE_DIR = PATHS.state_dir
-STATE_FILE = PATHS.precheck_state_file
-POSITIONS_FILE = PATHS.positions_file
-CONFIG_FILE = PATHS.config_file
-ENV_FILE = PATHS.env_file
+_BASE = "precheck"

-BASE_PRICE_MOVE_TRIGGER_PCT = 0.025
-BASE_PNL_TRIGGER_PCT = 0.03
-BASE_PORTFOLIO_MOVE_TRIGGER_PCT = 0.03
-BASE_CANDIDATE_SCORE_TRIGGER_RATIO = 1.15
-BASE_FORCE_ANALYSIS_AFTER_MINUTES = 180
-BASE_COOLDOWN_MINUTES = 45
-TOP_CANDIDATES = 10
-MIN_ACTIONABLE_USDT = 12.0
-MIN_REAL_POSITION_VALUE_USDT = 8.0
-BLACKLIST = {"USDC", "BUSD", "TUSD", "FDUSD", "USTC", "PAXG"}
-HARD_STOP_PCT = -0.08
-HARD_MOON_PCT = 0.25
-MIN_CHANGE_PCT = 1.0
-MAX_PRICE_CAP = None
-HARD_REASON_DEDUP_MINUTES = 15
-MAX_PENDING_TRIGGER_MINUTES = 30
-MAX_RUN_REQUEST_MINUTES = 20
+BASE_PRICE_MOVE_TRIGGER_PCT = get_user_config(f"{_BASE}.base_price_move_trigger_pct", 0.025)
+BASE_PNL_TRIGGER_PCT = get_user_config(f"{_BASE}.base_pnl_trigger_pct", 0.03)
+BASE_PORTFOLIO_MOVE_TRIGGER_PCT = get_user_config(f"{_BASE}.base_portfolio_move_trigger_pct", 0.03)
+BASE_CANDIDATE_SCORE_TRIGGER_RATIO = get_user_config(f"{_BASE}.base_candidate_score_trigger_ratio", 1.15)
+BASE_FORCE_ANALYSIS_AFTER_MINUTES = get_user_config(f"{_BASE}.base_force_analysis_after_minutes", 180)
+BASE_COOLDOWN_MINUTES = get_user_config(f"{_BASE}.base_cooldown_minutes", 45)
+TOP_CANDIDATES = get_user_config(f"{_BASE}.top_candidates", 10)
+MIN_ACTIONABLE_USDT = get_user_config(f"{_BASE}.min_actionable_usdt", 12.0)
+MIN_REAL_POSITION_VALUE_USDT = get_user_config(f"{_BASE}.min_real_position_value_usdt", 8.0)
+BLACKLIST = set(get_user_config(f"{_BASE}.blacklist", ["USDC", "BUSD", "TUSD", "FDUSD", "USTC", "PAXG"]))
+HARD_STOP_PCT = get_user_config(f"{_BASE}.hard_stop_pct", -0.08)
+HARD_MOON_PCT = get_user_config(f"{_BASE}.hard_moon_pct", 0.25)
+MIN_CHANGE_PCT = get_user_config(f"{_BASE}.min_change_pct", 1.0)
+MAX_PRICE_CAP = get_user_config(f"{_BASE}.max_price_cap", None)
+HARD_REASON_DEDUP_MINUTES = get_user_config(f"{_BASE}.hard_reason_dedup_minutes", 15)
+MAX_PENDING_TRIGGER_MINUTES = get_user_config(f"{_BASE}.max_pending_trigger_minutes", 30)
+MAX_RUN_REQUEST_MINUTES = get_user_config(f"{_BASE}.max_run_request_minutes", 20)
--- a/src/coinhunter/services/precheck_core.py
+++ b/src/coinhunter/services/precheck_core.py
@@ -18,14 +18,19 @@ from __future__ import annotations

 from importlib import import_module

+from ..runtime import get_runtime_paths
+
+_PATH_ALIASES = {
+    "PATHS": lambda: get_runtime_paths(),
+    "BASE_DIR": lambda: get_runtime_paths().root,
+    "STATE_DIR": lambda: get_runtime_paths().state_dir,
+    "STATE_FILE": lambda: get_runtime_paths().precheck_state_file,
+    "POSITIONS_FILE": lambda: get_runtime_paths().positions_file,
+    "CONFIG_FILE": lambda: get_runtime_paths().config_file,
+    "ENV_FILE": lambda: get_runtime_paths().env_file,
+}
+
 _MODULE_MAP = {
-    "PATHS": ".precheck_constants",
-    "BASE_DIR": ".precheck_constants",
-    "STATE_DIR": ".precheck_constants",
-    "STATE_FILE": ".precheck_constants",
-    "POSITIONS_FILE": ".precheck_constants",
-    "CONFIG_FILE": ".precheck_constants",
-    "ENV_FILE": ".precheck_constants",
    "BASE_PRICE_MOVE_TRIGGER_PCT": ".precheck_constants",
    "BASE_PNL_TRIGGER_PCT": ".precheck_constants",
    "BASE_PORTFOLIO_MOVE_TRIGGER_PCT": ".precheck_constants",
@@ -74,10 +79,12 @@ _MODULE_MAP = {
    "analyze_trigger": ".trigger_analyzer",
 }

-__all__ = sorted(set(_MODULE_MAP) | {"main"})
+__all__ = sorted(set(_MODULE_MAP) | set(_PATH_ALIASES) | {"main"})


 def __getattr__(name: str):
+    if name in _PATH_ALIASES:
+        return _PATH_ALIASES[name]()
    if name not in _MODULE_MAP:
        raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
    module_name = _MODULE_MAP[name]
@@ -90,8 +97,9 @@ def __dir__():


 def main():
-    from .precheck_service import run as _run_service
    import sys
+
+    from .precheck_service import run as _run_service
    return _run_service(sys.argv[1:])


--- a/src/coinhunter/services/precheck_service.py
+++ b/src/coinhunter/services/precheck_service.py
@@ -3,6 +3,8 @@
 from __future__ import annotations

 import json
+
+__all__ = ["run"]
 import sys

 from . import precheck_analysis, precheck_snapshot, precheck_state
@@ -19,12 +21,24 @@ def run(argv: list[str] | None = None) -> int:
        return 0

    try:
-        state = precheck_state.sanitize_state_for_stale_triggers(precheck_state.load_state())
-        snapshot = precheck_snapshot.build_snapshot()
-        analysis = precheck_analysis.analyze_trigger(snapshot, state)
-        precheck_state.save_state(precheck_state.update_state_after_observation(state, snapshot, analysis))
-        print(json.dumps(analysis, ensure_ascii=False, indent=2))
+        captured = {}
+
+        def _modifier(state):
+            state = precheck_state.sanitize_state_for_stale_triggers(state)
+            snapshot = precheck_snapshot.build_snapshot()
+            analysis = precheck_analysis.analyze_trigger(snapshot, state)
+            new_state = precheck_state.update_state_after_observation(state, snapshot, analysis)
+            state.clear()
+            state.update(new_state)
+            captured["analysis"] = analysis
+            return state
+
+        precheck_state.modify_state(_modifier)
+        result = {"ok": True, **captured["analysis"]}
+        print(json.dumps(result, ensure_ascii=False, indent=2))
        return 0
    except Exception as exc:
-        print(json.dumps(precheck_analysis.build_failure_payload(exc), ensure_ascii=False, indent=2))
-        return 0
+        payload = precheck_analysis.build_failure_payload(exc)
+        result = {"ok": False, **payload}
+        print(json.dumps(result, ensure_ascii=False, indent=2))
+        return 1
--- a/src/coinhunter/services/precheck_snapshot.py
+++ b/src/coinhunter/services/precheck_snapshot.py
@@ -2,4 +2,3 @@

 from __future__ import annotations

-from .snapshot_builder import build_snapshot
--- a/src/coinhunter/services/precheck_state.py
+++ b/src/coinhunter/services/precheck_state.py
@@ -6,32 +6,36 @@ import json

 from .state_manager import (
    load_state,
-    sanitize_state_for_stale_triggers,
-    save_state,
-    update_state_after_observation,
+    modify_state,
 )
 from .time_utils import utc_iso


 def mark_run_requested(note: str = "") -> dict:
+    def _modifier(state):
+        state["run_requested_at"] = utc_iso()
+        state["run_request_note"] = note
+        return state
+
+    modify_state(_modifier)
    state = load_state()
-    state["run_requested_at"] = utc_iso()
-    state["run_request_note"] = note
-    save_state(state)
    payload = {"ok": True, "run_requested_at": state["run_requested_at"], "note": note}
    print(json.dumps(payload, ensure_ascii=False))
    return payload


 def ack_analysis(note: str = "") -> dict:
+    def _modifier(state):
+        state["last_deep_analysis_at"] = utc_iso()
+        state["pending_trigger"] = False
+        state["pending_reasons"] = []
+        state["last_ack_note"] = note
+        state.pop("run_requested_at", None)
+        state.pop("run_request_note", None)
+        return state
+
+    modify_state(_modifier)
    state = load_state()
-    state["last_deep_analysis_at"] = utc_iso()
-    state["pending_trigger"] = False
-    state["pending_reasons"] = []
-    state["last_ack_note"] = note
-    state.pop("run_requested_at", None)
-    state.pop("run_request_note", None)
-    save_state(state)
    payload = {"ok": True, "acked_at": state["last_deep_analysis_at"], "note": note}
    print(json.dumps(payload, ensure_ascii=False))
    return payload
--- a/src/coinhunter/services/review_service.py
+++ b/src/coinhunter/services/review_service.py
@@ -0,0 +1,292 @@
+"""Review generation service for CoinHunter."""
+
+from __future__ import annotations
+
+import json
+
+__all__ = [
+    "ensure_review_dir",
+    "norm_symbol",
+    "fetch_current_price",
+    "analyze_trade",
+    "analyze_hold_passes",
+    "analyze_cash_misses",
+    "generate_review",
+    "save_review",
+    "print_review",
+]
+from datetime import datetime, timedelta, timezone
+from pathlib import Path
+
+from ..logger import get_logs_last_n_hours
+from ..runtime import get_runtime_paths
+from .exchange_service import get_exchange
+
+CST = timezone(timedelta(hours=8))
+
+
+def _paths():
+    return get_runtime_paths()
+
+
+def _review_dir() -> Path:
+    return _paths().reviews_dir  # type: ignore[no-any-return]
+
+
+def _exchange():
+    return get_exchange()
+
+
+def ensure_review_dir():
+    _review_dir().mkdir(parents=True, exist_ok=True)
+
+
+def norm_symbol(symbol: str) -> str:
+    s = symbol.upper().replace("-", "").replace("_", "")
+    if "/" in s:
+        return s
+    if s.endswith("USDT"):
+        return s[:-4] + "/USDT"
+    return s
+
+
+def fetch_current_price(ex, symbol: str):
+    try:
+        return float(ex.fetch_ticker(norm_symbol(symbol))["last"])
+    except Exception:
+        return None
+
+
+def analyze_trade(trade: dict, ex) -> dict:
+    symbol = trade.get("symbol")
+    price = trade.get("price")
+    action = trade.get("action", "")
+    current_price = fetch_current_price(ex, symbol) if symbol else None
+    pnl_estimate = None
+    outcome = "neutral"
+    if price and current_price and symbol:
+        change_pct = (current_price - float(price)) / float(price) * 100
+        if action == "BUY":
+            pnl_estimate = round(change_pct, 2)
+            outcome = "good" if change_pct > 2 else "bad" if change_pct < -2 else "neutral"
+        elif action == "SELL_ALL":
+            pnl_estimate = round(-change_pct, 2)
+            outcome = "good" if change_pct < -2 else "missed" if change_pct > 2 else "neutral"
+    return {
+        "timestamp": trade.get("timestamp"),
+        "symbol": symbol,
+        "action": action,
+        "decision_id": trade.get("decision_id"),
+        "execution_price": price,
+        "current_price": current_price,
+        "pnl_estimate_pct": pnl_estimate,
+        "outcome_assessment": outcome,
+    }
+
+
+def analyze_hold_passes(decisions: list, ex) -> list:
+    misses = []
+    for d in decisions:
+        if d.get("decision") != "HOLD":
+            continue
+        analysis = d.get("analysis")
+        if not isinstance(analysis, dict):
+            continue
+        opportunities = analysis.get("opportunities_evaluated", [])
+        market_snapshot = d.get("market_snapshot", {})
+        if not opportunities or not market_snapshot:
+            continue
+        for op in opportunities:
+            verdict = op.get("verdict", "")
+            if "PASS" not in verdict and "pass" not in verdict:
+                continue
+            symbol = op.get("symbol", "")
+            snap = market_snapshot.get(symbol) or market_snapshot.get(symbol.replace("/", ""))
+            if not snap:
+                continue
+            decision_price = None
+            if isinstance(snap, dict):
+                decision_price = float(snap.get("lastPrice", 0)) or float(snap.get("last", 0))
+            elif isinstance(snap, (int, float, str)):
+                decision_price = float(snap)
+            if not decision_price:
+                continue
+            current_price = fetch_current_price(ex, symbol)
+            if not current_price:
+                continue
+            change_pct = (current_price - decision_price) / decision_price * 100
+            if change_pct > 3:
+                misses.append({
+                    "timestamp": d.get("timestamp"),
+                    "symbol": symbol,
+                    "decision_price": round(decision_price, 8),
+                    "current_price": round(current_price, 8),
+                    "change_pct": round(change_pct, 2),
+                    "verdict_snippet": verdict[:80],
+                })
+    return misses
+
+
+def analyze_cash_misses(decisions: list, ex) -> list:
+    misses = []
+    watchlist = set()
+    for d in decisions:
+        snap = d.get("market_snapshot", {})
+        if isinstance(snap, dict):
+            for k in snap.keys():
+                if k.endswith("USDT"):
+                    watchlist.add(k)
+    for d in decisions:
+        ts = d.get("timestamp")
+        balances = d.get("balances") or d.get("balances_before", {})
+        if not balances:
+            continue
+        total = sum(float(v) if isinstance(v, (int, float, str)) else 0 for v in balances.values())
+        usdt = float(balances.get("USDT", 0))
+        if total == 0 or (usdt / total) < 0.9:
+            continue
+        snap = d.get("market_snapshot", {})
+        if not isinstance(snap, dict):
+            continue
+        for symbol, data in snap.items():
+            if not symbol.endswith("USDT"):
+                continue
+            decision_price = None
+            if isinstance(data, dict):
+                decision_price = float(data.get("lastPrice", 0)) or float(data.get("last", 0))
+            elif isinstance(data, (int, float, str)):
+                decision_price = float(data)
+            if not decision_price:
+                continue
+            current_price = fetch_current_price(ex, symbol)
+            if not current_price:
+                continue
+            change_pct = (current_price - decision_price) / decision_price * 100
+            if change_pct > 5:
+                misses.append({
+                    "timestamp": ts,
+                    "symbol": symbol,
+                    "decision_price": round(decision_price, 8),
+                    "current_price": round(current_price, 8),
+                    "change_pct": round(change_pct, 2),
+                })
+    seen: dict[str, dict] = {}
+    for m in misses:
+        sym = m["symbol"]
+        if sym not in seen or m["change_pct"] > seen[sym]["change_pct"]:
+            seen[sym] = m
+    return list(seen.values())
+
+
+def generate_review(hours: int = 1) -> dict:
+    decisions = get_logs_last_n_hours("decisions", hours)
+    trades = get_logs_last_n_hours("trades", hours)
+    errors = get_logs_last_n_hours("errors", hours)
+
+    review: dict = {
+        "review_period_hours": hours,
+        "review_timestamp": datetime.now(CST).isoformat(),
+        "total_decisions": len(decisions),
+        "total_trades": len(trades),
+        "total_errors": len(errors),
+        "decision_quality": [],
+        "stats": {},
+        "insights": [],
+        "recommendations": [],
+    }
+
+    if not decisions and not trades:
+        review["insights"].append("No decisions or trades in this period")
+        return review
+
+    ex = get_exchange()
+    outcomes = {"good": 0, "neutral": 0, "bad": 0, "missed": 0}
+    pnl_samples = []
+
+    for trade in trades:
+        analysis = analyze_trade(trade, ex)
+        review["decision_quality"].append(analysis)
+        outcomes[analysis["outcome_assessment"]] += 1
+        if analysis["pnl_estimate_pct"] is not None:
+            pnl_samples.append(analysis["pnl_estimate_pct"])
+
+    hold_pass_misses = analyze_hold_passes(decisions, ex)
+    cash_misses = analyze_cash_misses(decisions, ex)
+    total_missed = outcomes["missed"] + len(hold_pass_misses) + len(cash_misses)
+
+    review["stats"] = {
+        "good_decisions": outcomes["good"],
+        "neutral_decisions": outcomes["neutral"],
+        "bad_decisions": outcomes["bad"],
+        "missed_opportunities": total_missed,
+        "missed_sell_all": outcomes["missed"],
+        "missed_hold_passes": len(hold_pass_misses),
+        "missed_cash_sits": len(cash_misses),
+        "avg_estimated_edge_pct": round(sum(pnl_samples) / len(pnl_samples), 2) if pnl_samples else None,
+    }
+
+    if errors:
+        review["insights"].append(f"{len(errors)} execution/system errors this period; robustness needs attention")
+    if outcomes["bad"] > outcomes["good"]:
+        review["insights"].append("Recent trade quality is weak; consider lowering frequency or raising entry threshold")
+    if total_missed > 0:
+        parts = []
+        if outcomes["missed"]:
+            parts.append(f"sold then rallied {outcomes['missed']} times")
+        if hold_pass_misses:
+            parts.append(f"missed after PASS {len(hold_pass_misses)} times")
+        if cash_misses:
+            parts.append(f"missed while sitting in cash {len(cash_misses)} times")
+        review["insights"].append("Opportunities missed: " + ", ".join(parts) + "; consider relaxing trend-following or entry conditions")
+    if outcomes["good"] >= max(1, outcomes["bad"] + total_missed):
+        review["insights"].append("Recent decisions are generally acceptable")
+    if not trades and decisions:
+        review["insights"].append("Decisions without trades; may be due to waiting on sidelines, minimum notional limits, or execution interception")
+    if len(trades) < len(decisions) * 0.1 and decisions:
+        review["insights"].append("Many decisions did not convert to trades; check if minimum notional/step-size/fee buffer thresholds are too high")
+    if hold_pass_misses:
+        for m in hold_pass_misses[:3]:
+            review["insights"].append(f"PASS'd {m['symbol']} during HOLD, then it rose {m['change_pct']}%")
+    if cash_misses:
+        for m in cash_misses[:3]:
+            review["insights"].append(f"{m['symbol']} rose {m['change_pct']}% while portfolio was mostly USDT")
+
+    review["recommendations"] = [
+        "Check whether minimum-notional/precision rejections are blocking small-capital execution",
+        "If estimated edge is negative for two consecutive review periods, reduce rebalancing frequency next hour",
+        "If error logs are increasing, prioritize defensive mode (hold more USDT)",
+    ]
+    return review
+
+
+def save_review(review: dict) -> str:
+    ensure_review_dir()
+    ts = datetime.now(CST).strftime("%Y%m%d_%H%M%S")
+    path = _review_dir() / f"review_{ts}.json"
+    path.write_text(json.dumps(review, indent=2, ensure_ascii=False), encoding="utf-8")
+    return str(path)
+
+
+def print_review(review: dict):
+    print("=" * 50)
+    print("📊 Coin Hunter Review Report")
+    print(f"Review time: {review['review_timestamp']}")
+    print(f"Period: last {review['review_period_hours']} hours")
+    print(f"Total decisions: {review['total_decisions']} | Total trades: {review['total_trades']} | Total errors: {review['total_errors']}")
+    stats = review.get("stats", {})
+    print("\nDecision quality:")
+    print(f"  ✓ Good: {stats.get('good_decisions', 0)}")
+    print(f"  ○ Neutral: {stats.get('neutral_decisions', 0)}")
+    print(f"  ✗ Bad: {stats.get('bad_decisions', 0)}")
+    print(f"  ↗ Missed opportunities: {stats.get('missed_opportunities', 0)}")
+    if stats.get("avg_estimated_edge_pct") is not None:
+        print(f"  Avg estimated edge: {stats['avg_estimated_edge_pct']}%")
+    if review.get("insights"):
+        print("\n💡 Insights:")
+        for item in review["insights"]:
+            print(f"  • {item}")
+    if review.get("recommendations"):
+        print("\n🔧 Recommendations:")
+        for item in review["recommendations"]:
+            print(f"  • {item}")
+    print("=" * 50)
--- a/src/coinhunter/services/smart_executor_parser.py
+++ b/src/coinhunter/services/smart_executor_parser.py
@@ -1,6 +1,15 @@
 """CLI parser and legacy argument normalization for smart executor."""
 import argparse

+__all__ = [
+    "COMMAND_CANONICAL",
+    "add_shared_options",
+    "build_parser",
+    "normalize_legacy_argv",
+    "parse_cli_args",
+    "cli_action_args",
+]
+

 COMMAND_CANONICAL = {
    "bal": "balances",
@@ -16,6 +25,10 @@ COMMAND_CANONICAL = {
    "sell_all": "sell-all",
    "rotate": "rebalance",
    "rebalance": "rebalance",
+    "orders": "orders",
+    "cancel": "cancel",
+    "order-status": "order-status",
+    "order_status": "order-status",
 }


@@ -41,25 +54,32 @@ def build_parser() -> argparse.ArgumentParser:
            "  hold                Record a hold decision without trading\n"
            "  buy SYMBOL USDT     Buy a symbol using a USDT notional amount\n"
            "  flat SYMBOL         Exit an entire symbol position\n"
-            "  rotate FROM TO      Rotate exposure from one symbol into another\n\n"
+            "  rotate FROM TO      Rotate exposure from one symbol into another\n"
+            "  orders              List open spot orders\n"
+            "  order-status SYMBOL ORDER_ID  Get status of a specific order\n"
+            "  cancel SYMBOL [ORDER_ID]      Cancel an open order (cancels newest if ORDER_ID omitted)\n\n"
            "Examples:\n"
            "  coinhunter exec bal\n"
            "  coinhunter exec overview\n"
            "  coinhunter exec hold\n"
            "  coinhunter exec buy ENJUSDT 100\n"
            "  coinhunter exec flat ENJUSDT --dry-run\n"
-            "  coinhunter exec rotate PEPEUSDT ETHUSDT\n\n"
+            "  coinhunter exec rotate PEPEUSDT ETHUSDT\n"
+            "  coinhunter exec orders\n"
+            "  coinhunter exec order-status ENJUSDT 123456\n"
+            "  coinhunter exec cancel ENJUSDT 123456\n\n"
            "Legacy forms remain supported for backward compatibility:\n"
            "  balances, balance -> bal\n"
            "  acct, status -> overview\n"
            "  sell-all, sell_all -> flat\n"
            "  rebalance -> rotate\n"
+            "  order_status -> order-status\n"
            "  HOLD / BUY / SELL_ALL / REBALANCE via --decision are still accepted\n"
        ),
    )
    subparsers = parser.add_subparsers(
        dest="command",
-        metavar="{bal,overview,hold,buy,flat,rotate,...}",
+        metavar="{bal,overview,hold,buy,flat,rotate,orders,order-status,cancel,...}",
    )

    subparsers.add_parser("bal", parents=[shared], help="Preferred: print live balances as stable JSON")
@@ -77,6 +97,16 @@ def build_parser() -> argparse.ArgumentParser:
    rebalance.add_argument("from_symbol")
    rebalance.add_argument("to_symbol")

+    subparsers.add_parser("orders", parents=[shared], help="List open spot orders")
+
+    order_status = subparsers.add_parser("order-status", parents=[shared], help="Get status of a specific order")
+    order_status.add_argument("symbol")
+    order_status.add_argument("order_id")
+
+    cancel = subparsers.add_parser("cancel", parents=[shared], help="Cancel an open order")
+    cancel.add_argument("symbol")
+    cancel.add_argument("order_id", nargs="?")
+
    subparsers.add_parser("balances", parents=[shared], help=argparse.SUPPRESS)
    subparsers.add_parser("balance", parents=[shared], help=argparse.SUPPRESS)
    subparsers.add_parser("acct", parents=[shared], help=argparse.SUPPRESS)
@@ -91,6 +121,10 @@ def build_parser() -> argparse.ArgumentParser:
    rebalance_legacy.add_argument("from_symbol")
    rebalance_legacy.add_argument("to_symbol")

+    order_status_legacy = subparsers.add_parser("order_status", parents=[shared], help=argparse.SUPPRESS)
+    order_status_legacy.add_argument("symbol")
+    order_status_legacy.add_argument("order_id")
+
    subparsers._choices_actions = [
        action
        for action in subparsers._choices_actions
@@ -126,7 +160,13 @@ def normalize_legacy_argv(argv: list[str]) -> list[str]:
        "STATUS": ["status"],
        "status": ["status"],
        "OVERVIEW": ["status"],
-        "overview": ["status"],
+        "ORDERS": ["orders"],
+        "orders": ["orders"],
+        "CANCEL": ["cancel"],
+        "cancel": ["cancel"],
+        "ORDER_STATUS": ["order-status"],
+        "order_status": ["order-status"],
+        "order-status": ["order-status"],
    }

    has_legacy_flag = any(t.startswith("--decision") for t in argv)
@@ -166,18 +206,18 @@ def normalize_legacy_argv(argv: list[str]) -> list[str]:
                rebuilt += ["hold"]
            elif decision == "SELL_ALL":
                if not ns.symbol:
-                    raise RuntimeError("旧式 --decision SELL_ALL 需要搭配 --symbol")
+                    raise RuntimeError("Legacy --decision SELL_ALL requires --symbol")
                rebuilt += ["sell-all", ns.symbol]
            elif decision == "BUY":
                if not ns.symbol or ns.amount_usdt is None:
-                    raise RuntimeError("旧式 --decision BUY 需要 --symbol 和 --amount-usdt")
+                    raise RuntimeError("Legacy --decision BUY requires --symbol and --amount-usdt")
                rebuilt += ["buy", ns.symbol, str(ns.amount_usdt)]
            elif decision == "REBALANCE":
                if not ns.from_symbol or not ns.to_symbol:
-                    raise RuntimeError("旧式 --decision REBALANCE 需要 --from-symbol 和 --to-symbol")
+                    raise RuntimeError("Legacy --decision REBALANCE requires --from-symbol and --to-symbol")
                rebuilt += ["rebalance", ns.from_symbol, ns.to_symbol]
            else:
-                raise RuntimeError(f"不支持的旧式 decision: {decision}")
+                raise RuntimeError(f"Unsupported legacy decision: {decision}")

            return rebuilt + unknown

@@ -202,4 +242,8 @@ def cli_action_args(args, action: str) -> list[str]:
        return [args.symbol, str(args.amount_usdt)]
    if action == "rebalance":
        return [args.from_symbol, args.to_symbol]
+    if action == "order_status":
+        return [args.symbol, args.order_id]
+    if action == "cancel":
+        return [args.symbol, args.order_id] if args.order_id else [args.symbol]
    return []
--- a/src/coinhunter/services/smart_executor_service.py
+++ b/src/coinhunter/services/smart_executor_service.py
@@ -3,21 +3,27 @@
 from __future__ import annotations

 import os
+
+__all__ = ["run"]
 import sys

 from ..logger import log_decision, log_error
-from .exchange_service import fetch_balances, build_market_snapshot
+from .exchange_service import build_market_snapshot, fetch_balances
 from .execution_state import default_decision_id, get_execution_state, record_execution_state
 from .portfolio_service import load_positions
-from .smart_executor_parser import parse_cli_args, cli_action_args
-from .trade_common import is_dry_run, log, set_dry_run, bj_now_iso
+from .smart_executor_parser import cli_action_args, parse_cli_args
+from .trade_common import bj_now_iso, log, set_dry_run
 from .trade_execution import (
-    command_balances,
-    command_status,
-    build_decision_context,
-    action_sell_all,
    action_buy,
    action_rebalance,
+    action_sell_all,
+    build_decision_context,
+    command_balances,
+    command_cancel,
+    command_order_status,
+    command_orders,
+    command_status,
+    print_json,
 )


@@ -36,9 +42,9 @@ def run(argv: list[str] | None = None) -> int:
        set_dry_run(True)

    previous = get_execution_state(decision_id)
-    read_only_action = action in {"balance", "balances", "status"}
+    read_only_action = action in {"balance", "balances", "status", "orders", "order_status", "cancel"}
    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")
+        log(f"⚠️ decision_id={decision_id} already executed successfully, skipping duplicate")
        return 0

    try:
@@ -48,8 +54,14 @@ def run(argv: list[str] | None = None) -> int:
        if read_only_action:
            if action in {"balance", "balances"}:
                command_balances(ex)
-            else:
+            elif action == "status":
                command_status(ex)
+            elif action == "orders":
+                command_orders(ex)
+            elif action == "order_status":
+                command_order_status(ex, args.symbol, args.order_id)
+            elif action == "cancel":
+                command_cancel(ex, args.symbol, getattr(args, "order_id", None))
            return 0

        decision_context = build_decision_context(ex, action, argv_tail, decision_id)
@@ -88,10 +100,10 @@ def run(argv: list[str] | None = None) -> int:
                    "execution_result": {"status": "hold"},
                }
            )
-            log("😴 决策: 持续持有，无操作")
+            log("😴 Decision: hold, no action")
            result = {"status": "hold"}
        else:
-            raise RuntimeError(f"未知动作: {action}；请运行 --help 查看正确 CLI 用法")
+            raise RuntimeError(f"Unknown action: {action}; run --help for valid CLI usage")

        record_execution_state(
            decision_id,
@@ -103,7 +115,9 @@ def run(argv: list[str] | None = None) -> int:
                "result": result,
            },
        )
-        log(f"✅ 执行完成 decision_id={decision_id}")
+        if not read_only_action:
+            print_json({"ok": True, "decision_id": decision_id, "action": action, "result": result})
+        log(f"Execution completed decision_id={decision_id}")
        return 0

    except Exception as exc:
@@ -124,5 +138,5 @@ def run(argv: list[str] | None = None) -> int:
            action=action,
            args=argv_tail,
        )
-        log(f"❌ 执行失败: {exc}")
+        log(f"❌ Execution failed: {exc}")
        return 1
--- a/src/coinhunter/services/snapshot_builder.py
+++ b/src/coinhunter/services/snapshot_builder.py
@@ -5,7 +5,7 @@ from __future__ import annotations
 from .candidate_scoring import top_candidates_from_tickers
 from .data_utils import norm_symbol, stable_hash, to_float
 from .market_data import enrich_candidates_and_positions, get_exchange, regime_from_pct
-from .precheck_constants import MIN_ACTIONABLE_USDT, MIN_REAL_POSITION_VALUE_USDT
+from .precheck_constants import MIN_REAL_POSITION_VALUE_USDT
 from .state_manager import load_config, load_positions
 from .time_utils import get_local_now, utc_iso

--- a/src/coinhunter/services/state_manager.py
+++ b/src/coinhunter/services/state_manager.py
@@ -4,34 +4,58 @@ from __future__ import annotations

 from datetime import timedelta

-from ..runtime import load_env_file
+__all__ = [
+    "load_env",
+    "load_positions",
+    "load_state",
+    "modify_state",
+    "load_config",
+    "clear_run_request_fields",
+    "sanitize_state_for_stale_triggers",
+    "save_state",
+    "update_state_after_observation",
+]
+
+from ..runtime import get_runtime_paths, load_env_file
 from .data_utils import load_json
+from .file_utils import load_json_locked, read_modify_write_json, save_json_locked
 from .precheck_constants import (
-    CONFIG_FILE,
-    ENV_FILE,
    MAX_PENDING_TRIGGER_MINUTES,
    MAX_RUN_REQUEST_MINUTES,
-    POSITIONS_FILE,
-    STATE_DIR,
-    STATE_FILE,
 )
 from .time_utils import parse_ts, utc_iso, utc_now


+def _paths():
+    return get_runtime_paths()
+
+
 def load_env() -> None:
-    load_env_file()
+    load_env_file(_paths())


 def load_positions():
-    return load_json(POSITIONS_FILE, {}).get("positions", [])
+    return load_json(_paths().positions_file, {}).get("positions", [])


 def load_state():
-    return load_json(STATE_FILE, {})
+    paths = _paths()
+    return load_json_locked(paths.precheck_state_file, paths.precheck_state_lock, {})
+
+
+def modify_state(modifier):
+    """Atomic read-modify-write for precheck state."""
+    paths = _paths()
+    read_modify_write_json(
+        paths.precheck_state_file,
+        paths.precheck_state_lock,
+        {},
+        modifier,
+    )


 def load_config():
-    return load_json(CONFIG_FILE, {})
+    return load_json(_paths().config_file, {})


 def clear_run_request_fields(state: dict):
@@ -58,14 +82,14 @@ def sanitize_state_for_stale_triggers(state: dict):
            )
            pending_trigger = False
        notes.append(
-            f"自动清理已完成的 run_requested 标记：最近深度分析时间 {last_deep_analysis_at.isoformat()} >= 请求时间 {run_requested_at.isoformat()}"
+            f"Auto-cleared completed run_requested marker: last_deep_analysis_at {last_deep_analysis_at.isoformat()} >= run_requested_at {run_requested_at.isoformat()}"
        )
        run_requested_at = None

    if run_requested_at and now - run_requested_at > timedelta(minutes=MAX_RUN_REQUEST_MINUTES):
        clear_run_request_fields(sanitized)
        notes.append(
-            f"自动清理超时 run_requested 标记：已等待 {(now - run_requested_at).total_seconds() / 60:.1f} 分钟，超过 {MAX_RUN_REQUEST_MINUTES} 分钟"
+            f"Auto-cleared stale run_requested marker: waited {(now - run_requested_at).total_seconds() / 60:.1f} minutes, exceeding {MAX_RUN_REQUEST_MINUTES} minutes"
        )
        run_requested_at = None

@@ -78,7 +102,7 @@ def sanitize_state_for_stale_triggers(state: dict):
            f"{(now - pending_anchor).total_seconds() / 60:.1f} minutes"
        )
        notes.append(
-            f"自动解除 pending_trigger：触发状态已悬挂 {(now - pending_anchor).total_seconds() / 60:.1f} 分钟，超过 {MAX_PENDING_TRIGGER_MINUTES} 分钟"
+            f"Auto-recovered stale pending_trigger: trigger was dangling for {(now - pending_anchor).total_seconds() / 60:.1f} minutes, exceeding {MAX_PENDING_TRIGGER_MINUTES} minutes"
        )

    sanitized["_stale_recovery_notes"] = notes
@@ -86,12 +110,16 @@ def sanitize_state_for_stale_triggers(state: dict):


 def save_state(state: dict):
-    import json

-    STATE_DIR.mkdir(parents=True, exist_ok=True)
+    paths = _paths()
+    paths.state_dir.mkdir(parents=True, exist_ok=True)
    state_to_save = dict(state)
    state_to_save.pop("_stale_recovery_notes", None)
-    STATE_FILE.write_text(json.dumps(state_to_save, indent=2, ensure_ascii=False), encoding="utf-8")
+    save_json_locked(
+        paths.precheck_state_file,
+        paths.precheck_state_lock,
+        state_to_save,
+    )


 def update_state_after_observation(state: dict, snapshot: dict, analysis: dict):
@@ -123,6 +151,10 @@ def update_state_after_observation(state: dict, snapshot: dict, analysis: dict):
    for hr in analysis.get("hard_reasons", []):
        last_hard_reasons_at[hr] = snapshot["generated_at"]
    cutoff = utc_now() - timedelta(hours=24)
-    pruned = {k: v for k, v in last_hard_reasons_at.items() if parse_ts(v) and parse_ts(v) > cutoff}
+    pruned: dict[str, str] = {}
+    for k, v in last_hard_reasons_at.items():
+        ts = parse_ts(v)
+        if ts and ts > cutoff:
+            pruned[k] = v
    new_state["last_hard_reasons_at"] = pruned
    return new_state
--- a/src/coinhunter/services/time_utils.py
+++ b/src/coinhunter/services/time_utils.py
@@ -2,7 +2,7 @@

 from __future__ import annotations

-from datetime import datetime, timedelta, timezone
+from datetime import datetime, timezone
 from zoneinfo import ZoneInfo


--- a/src/coinhunter/services/trade_common.py
+++ b/src/coinhunter/services/trade_common.py
@@ -1,12 +1,15 @@
 """Common trade utilities (time, logging, constants)."""
 import os
-from datetime import datetime, timezone, timedelta
+import sys
+from datetime import datetime, timedelta, timezone
+
+from ..runtime import get_user_config

 CST = timezone(timedelta(hours=8))

 _DRY_RUN = {"value": os.getenv("DRY_RUN", "false").lower() == "true"}
-USDT_BUFFER_PCT = 0.03
-MIN_REMAINING_DUST_USDT = 1.0
+USDT_BUFFER_PCT = get_user_config("trading.usdt_buffer_pct", 0.03)
+MIN_REMAINING_DUST_USDT = get_user_config("trading.min_remaining_dust_usdt", 1.0)


 def is_dry_run() -> bool:
@@ -18,7 +21,7 @@ def set_dry_run(value: bool):


 def log(msg: str):
-    print(f"[{datetime.now(CST).strftime('%Y-%m-%d %H:%M:%S')} CST] {msg}")
+    print(f"[{datetime.now(CST).strftime('%Y-%m-%d %H:%M:%S')} CST] {msg}", file=sys.stderr)


 def bj_now_iso():
--- a/src/coinhunter/services/trade_execution.py
+++ b/src/coinhunter/services/trade_execution.py
@@ -1,17 +1,37 @@
 """Trade execution actions (buy, sell, rebalance, hold, status)."""
 import json

+__all__ = [
+    "print_json",
+    "build_decision_context",
+    "market_sell",
+    "market_buy",
+    "action_sell_all",
+    "action_buy",
+    "action_rebalance",
+    "command_status",
+    "command_balances",
+    "command_orders",
+    "command_order_status",
+    "command_cancel",
+]
+
 from ..logger import log_decision, log_trade
 from .exchange_service import (
+    build_market_snapshot,
    fetch_balances,
    norm_symbol,
-    storage_symbol,
-    build_market_snapshot,
    prepare_buy_quantity,
    prepare_sell_quantity,
+    storage_symbol,
 )
-from .portfolio_service import load_positions, save_positions, upsert_position, reconcile_positions_with_exchange
-from .trade_common import is_dry_run, USDT_BUFFER_PCT, log, bj_now_iso
+from .portfolio_service import (
+    load_positions,
+    reconcile_positions_with_exchange,
+    update_positions,
+    upsert_position,
+)
+from .trade_common import USDT_BUFFER_PCT, bj_now_iso, is_dry_run, log


 def print_json(payload: dict) -> None:
@@ -35,7 +55,7 @@ def build_decision_context(ex, action: str, argv_tail: list[str], decision_id: s
 def market_sell(ex, symbol: str, qty: float, decision_id: str):
    sym, qty, bid, est_cost = prepare_sell_quantity(ex, symbol, qty)
    if is_dry_run():
-        log(f"[DRY RUN] 卖出 {sym} 数量 {qty}")
+        log(f"[DRY RUN] SELL {sym} qty {qty}")
        return {"id": f"dry-sell-{decision_id}", "symbol": sym, "amount": qty, "price": bid, "cost": est_cost, "status": "closed"}
    order = ex.create_market_sell_order(sym, qty, params={"newClientOrderId": f"ch-{decision_id}-sell"})
    return order
@@ -44,7 +64,7 @@ def market_sell(ex, symbol: str, qty: float, decision_id: str):
 def market_buy(ex, symbol: str, amount_usdt: float, decision_id: str):
    sym, qty, ask, est_cost = prepare_buy_quantity(ex, symbol, amount_usdt)
    if is_dry_run():
-        log(f"[DRY RUN] 买入 {sym} 金额 ${est_cost:.4f} 数量 {qty}")
+        log(f"[DRY RUN] BUY {sym} amount ${est_cost:.4f} qty {qty}")
        return {"id": f"dry-buy-{decision_id}", "symbol": sym, "amount": qty, "price": ask, "cost": est_cost, "status": "closed"}
    order = ex.create_market_buy_order(sym, qty, params={"newClientOrderId": f"ch-{decision_id}-buy"})
    return order
@@ -55,13 +75,13 @@ def action_sell_all(ex, symbol: str, decision_id: str, decision_context: dict):
    base = norm_symbol(symbol).split("/")[0]
    qty = float(balances_before.get(base, 0))
    if qty <= 0:
-        raise RuntimeError(f"{base} 余额为0，无法卖出")
+        raise RuntimeError(f"{base} balance is zero, cannot sell")
    order = market_sell(ex, symbol, qty, decision_id)
-    positions_after, balances_after = (
-        reconcile_positions_with_exchange(ex, load_positions())
-        if not is_dry_run()
-        else (load_positions(), balances_before)
-    )
+    if is_dry_run():
+        positions_after = load_positions()
+        balances_after = balances_before
+    else:
+        positions_after, balances_after = reconcile_positions_with_exchange(ex)
    log_trade(
        "SELL_ALL",
        norm_symbol(symbol),
@@ -88,13 +108,12 @@ def action_sell_all(ex, symbol: str, decision_id: str, decision_context: dict):
    return order


-def action_buy(ex, symbol: str, amount_usdt: float, decision_id: str, decision_context: dict, simulated_usdt_balance: float = None):
+def action_buy(ex, symbol: str, amount_usdt: float, decision_id: str, decision_context: dict, simulated_usdt_balance: float | None = None):
    balances_before = fetch_balances(ex) if simulated_usdt_balance is None else {"USDT": simulated_usdt_balance}
    usdt = float(balances_before.get("USDT", 0))
    if usdt < amount_usdt:
-        raise RuntimeError(f"USDT 余额不足（${usdt:.4f} < ${amount_usdt:.4f}）")
+        raise RuntimeError(f"Insufficient USDT balance (${usdt:.4f} < ${amount_usdt:.4f})")
    order = market_buy(ex, symbol, amount_usdt, decision_id)
-    positions_existing = load_positions()
    sym_store = storage_symbol(symbol)
    price = float(order.get("price") or 0)
    qty = float(order.get("amount") or 0)
@@ -110,18 +129,13 @@ def action_buy(ex, symbol: str, amount_usdt: float, decision_id: str, decision_c
        "updated_at": bj_now_iso(),
        "note": "Smart executor entry",
    }
-    upsert_position(positions_existing, position)
    if is_dry_run():
        balances_after = balances_before
-        positions_after = positions_existing
+        positions_after = load_positions()
+        upsert_position(positions_after, position)
    else:
-        save_positions(positions_existing)
-        positions_after, balances_after = reconcile_positions_with_exchange(ex, positions_existing)
-        for p in positions_after:
-            if p["symbol"] == sym_store and price:
-                p["avg_cost"] = price
-                p["updated_at"] = bj_now_iso()
-        save_positions(positions_after)
+        update_positions(lambda p: upsert_position(p, position))
+        positions_after, balances_after = reconcile_positions_with_exchange(ex, [position])
    log_trade(
        "BUY",
        norm_symbol(symbol),
@@ -160,7 +174,7 @@ def action_rebalance(ex, from_symbol: str, to_symbol: str, decision_id: str, dec
        spend = usdt * (1 - USDT_BUFFER_PCT)
        simulated_usdt = None
    if spend < 5:
-        raise RuntimeError(f"卖出后 USDT ${spend:.4f} 不足，无法买入新币")
+        raise RuntimeError(f"USDT ${spend:.4f} insufficient after sell, cannot buy new token")
    buy_order = action_buy(ex, to_symbol, spend, decision_id + "b", decision_context, simulated_usdt_balance=simulated_usdt)
    return {"sell": sell_order, "buy": buy_order}

@@ -183,3 +197,47 @@ def command_balances(ex):
    payload = {"balances": balances}
    print_json(payload)
    return balances
+
+
+def command_orders(ex):
+    sym = None
+    try:
+        orders = ex.fetch_open_orders(symbol=sym) if sym else ex.fetch_open_orders()
+    except Exception as e:
+        raise RuntimeError(f"Failed to fetch open orders: {e}")
+    payload = {"orders": orders}
+    print_json(payload)
+    return orders
+
+
+def command_order_status(ex, symbol: str, order_id: str):
+    sym = norm_symbol(symbol)
+    try:
+        order = ex.fetch_order(order_id, sym)
+    except Exception as e:
+        raise RuntimeError(f"Failed to fetch order {order_id}: {e}")
+    payload = {"order": order}
+    print_json(payload)
+    return order
+
+
+def command_cancel(ex, symbol: str, order_id: str | None):
+    sym = norm_symbol(symbol)
+    if is_dry_run():
+        log(f"[DRY RUN] Would cancel order {order_id or '(newest)'} on {sym}")
+        return {"dry_run": True, "symbol": sym, "order_id": order_id}
+    if not order_id:
+        try:
+            open_orders = ex.fetch_open_orders(sym)
+        except Exception as e:
+            raise RuntimeError(f"Failed to fetch open orders for {sym}: {e}")
+        if not open_orders:
+            raise RuntimeError(f"No open orders to cancel for {sym}")
+        order_id = str(open_orders[-1]["id"])
+    try:
+        result = ex.cancel_order(order_id, sym)
+    except Exception as e:
+        raise RuntimeError(f"Failed to cancel order {order_id} on {sym}: {e}")
+    payload = {"cancelled": True, "symbol": sym, "order_id": order_id, "result": result}
+    print_json(payload)
+    return result
--- a/src/coinhunter/services/trigger_analyzer.py
+++ b/src/coinhunter/services/trigger_analyzer.py
@@ -10,11 +10,9 @@ from .precheck_constants import (
    HARD_MOON_PCT,
    HARD_REASON_DEDUP_MINUTES,
    HARD_STOP_PCT,
-    MAX_PENDING_TRIGGER_MINUTES,
-    MAX_RUN_REQUEST_MINUTES,
    MIN_REAL_POSITION_VALUE_USDT,
 )
-from .time_utils import parse_ts, utc_iso, utc_now
+from .time_utils import parse_ts, utc_now


 def analyze_trigger(snapshot: dict, state: dict):
@@ -51,36 +49,37 @@ def analyze_trigger(snapshot: dict, state: dict):
    if pending_trigger:
        reasons.append("pending-trigger-unacked")
        hard_reasons.append("pending-trigger-unacked")
-        details.append("上次已触发深度分析但尚未确认完成")
+        details.append("Previous deep analysis trigger has not been acknowledged yet")
    if run_requested_at:
-        details.append(f"外部门控已在 {run_requested_at.isoformat()} 请求运行分析任务")
+        details.append(f"External gate requested analysis at {run_requested_at.isoformat()}")

    if not last_deep_analysis_at:
        reasons.append("first-analysis")
        hard_reasons.append("first-analysis")
-        details.append("尚未记录过深度分析")
+        details.append("No deep analysis has been recorded yet")
    elif now - last_deep_analysis_at >= timedelta(minutes=force_minutes):
        reasons.append("stale-analysis")
        hard_reasons.append("stale-analysis")
-        details.append(f"距离上次深度分析已超过 {force_minutes} 分钟")
+        details.append(f"Time since last deep analysis exceeds {force_minutes} minutes")

    if last_positions_hash and snapshot["positions_hash"] != last_positions_hash:
        reasons.append("positions-changed")
        hard_reasons.append("positions-changed")
-        details.append("持仓结构发生变化")
+        details.append("Position structure has changed")

    if last_portfolio_value not in (None, 0):
-        portfolio_delta = abs(snapshot["portfolio_value_usdt"] - last_portfolio_value) / max(last_portfolio_value, 1e-9)
+        lpf = float(str(last_portfolio_value))
+        portfolio_delta = abs(snapshot["portfolio_value_usdt"] - lpf) / max(lpf, 1e-9)
        if portfolio_delta >= portfolio_trigger:
            if portfolio_delta >= 1.0:
                reasons.append("portfolio-extreme-move")
                hard_reasons.append("portfolio-extreme-move")
-                details.append(f"组合净值剧烈变化 {portfolio_delta:.1%}，超过 100%，视为硬触发")
+                details.append(f"Portfolio value moved extremely {portfolio_delta:.1%}, exceeding 100%, treated as hard trigger")
            else:
                reasons.append("portfolio-move")
                soft_reasons.append("portfolio-move")
                soft_score += 1.0
-                details.append(f"组合净值变化 {portfolio_delta:.1%}，阈值 {portfolio_trigger:.1%}")
+                details.append(f"Portfolio value moved {portfolio_delta:.1%}, threshold {portfolio_trigger:.1%}")

    for pos in snapshot["positions"]:
        symbol = pos["symbol"]
@@ -98,42 +97,42 @@ def analyze_trigger(snapshot: dict, state: dict):
                reasons.append(f"price-move:{symbol}")
                soft_reasons.append(f"price-move:{symbol}")
                soft_score += 1.0 if actionable_position else 0.4
-                details.append(f"{symbol} 价格变化 {price_move:.1%}，阈值 {price_trigger:.1%}")
+                details.append(f"{symbol} price moved {price_move:.1%}, threshold {price_trigger:.1%}")
        if cur_pnl is not None and prev_pnl is not None:
            pnl_move = abs(cur_pnl - prev_pnl)
            if pnl_move >= pnl_trigger:
                reasons.append(f"pnl-move:{symbol}")
                soft_reasons.append(f"pnl-move:{symbol}")
                soft_score += 1.0 if actionable_position else 0.4
-                details.append(f"{symbol} 盈亏变化 {pnl_move:.1%}，阈值 {pnl_trigger:.1%}")
+                details.append(f"{symbol} PnL moved {pnl_move:.1%}, threshold {pnl_trigger:.1%}")
        if cur_pnl is not None:
            stop_band = -0.06 if actionable_position else -0.12
            take_band = 0.14 if actionable_position else 0.25
            if cur_pnl <= stop_band or cur_pnl >= take_band:
                reasons.append(f"risk-band:{symbol}")
                hard_reasons.append(f"risk-band:{symbol}")
-                details.append(f"{symbol} 接近执行阈值，当前盈亏 {cur_pnl:.1%}")
+                details.append(f"{symbol} near execution threshold, current PnL {cur_pnl:.1%}")
            if cur_pnl <= HARD_STOP_PCT:
                reasons.append(f"hard-stop:{symbol}")
                hard_reasons.append(f"hard-stop:{symbol}")
-                details.append(f"{symbol} 盈亏超过 {HARD_STOP_PCT:.1%}，触发紧急硬触发")
+                details.append(f"{symbol} PnL exceeded {HARD_STOP_PCT:.1%}, emergency hard trigger")

    current_market = snapshot.get("market_regime", {})
    if last_market_regime:
        if current_market.get("btc_regime") != last_market_regime.get("btc_regime"):
            reasons.append("btc-regime-change")
            hard_reasons.append("btc-regime-change")
-            details.append(f"BTC 由 {last_market_regime.get('btc_regime')} 切换为 {current_market.get('btc_regime')}")
+            details.append(f"BTC regime changed from {last_market_regime.get('btc_regime')} to {current_market.get('btc_regime')}")
        if current_market.get("eth_regime") != last_market_regime.get("eth_regime"):
            reasons.append("eth-regime-change")
            hard_reasons.append("eth-regime-change")
-            details.append(f"ETH 由 {last_market_regime.get('eth_regime')} 切换为 {current_market.get('eth_regime')}")
+            details.append(f"ETH regime changed from {last_market_regime.get('eth_regime')} to {current_market.get('eth_regime')}")

    for cand in snapshot.get("top_candidates", []):
        if cand.get("change_24h_pct", 0) >= HARD_MOON_PCT * 100:
            reasons.append(f"hard-moon:{cand['symbol']}")
            hard_reasons.append(f"hard-moon:{cand['symbol']}")
-            details.append(f"候选币 {cand['symbol']} 24h 涨幅 {cand['change_24h_pct']:.1f}%，触发强势硬触发")
+            details.append(f"Candidate {cand['symbol']} 24h change {cand['change_24h_pct']:.1f}%, hard moon trigger")

    candidate_weight = _candidate_weight(snapshot, profile)

@@ -151,7 +150,7 @@ def analyze_trigger(snapshot: dict, state: dict):
                soft_reasons.append(f"new-leader-{band}:{cur_leader['symbol']}")
                soft_score += candidate_weight * 0.7
                details.append(
-                    f"{band} 层新榜首 {cur_leader['symbol']} 替代 {prev_leader['symbol']}，score 比例 {score_ratio:.2f}"
+                    f"{band} tier new leader {cur_leader['symbol']} replaced {prev_leader['symbol']}, score ratio {score_ratio:.2f}"
                )

    current_leader = snapshot.get("top_candidates", [{}])[0] if snapshot.get("top_candidates") else None
@@ -163,13 +162,13 @@ def analyze_trigger(snapshot: dict, state: dict):
                soft_reasons.append("new-leader")
                soft_score += candidate_weight
                details.append(
-                    f"新候选币 {current_leader.get('symbol')} 领先上次榜首，score 比例 {score_ratio:.2f}，阈值 {candidate_ratio_trigger:.2f}"
+                    f"New candidate {current_leader.get('symbol')} leads previous top, score ratio {score_ratio:.2f}, threshold {candidate_ratio_trigger:.2f}"
                )
    elif current_leader and not last_top_candidate:
        reasons.append("candidate-leader-init")
        soft_reasons.append("candidate-leader-init")
        soft_score += candidate_weight
-        details.append(f"首次记录候选榜首 {current_leader.get('symbol')}")
+        details.append(f"First recorded candidate leader {current_leader.get('symbol')}")

    def _signal_delta() -> float:
        delta = 0.0
@@ -204,7 +203,8 @@ def analyze_trigger(snapshot: dict, state: dict):
            if current_market.get("eth_regime") != last_market_regime.get("eth_regime"):
                delta += 1.5
        if last_portfolio_value not in (None, 0):
-            portfolio_delta = abs(snapshot["portfolio_value_usdt"] - last_portfolio_value) / max(last_portfolio_value, 1e-9)
+            lpf = float(str(last_portfolio_value))
+            portfolio_delta = abs(snapshot["portfolio_value_usdt"] - lpf) / max(lpf, 1e-9)
            if portfolio_delta >= 0.05:
                delta += 1.0
        last_trigger_hard_types = {r.split(":")[0] for r in (state.get("last_trigger_hard_reasons") or [])}
@@ -227,41 +227,41 @@ def analyze_trigger(snapshot: dict, state: dict):
        last_at = parse_ts(last_hard_reasons_at.get(hr))
        if last_at and now - last_at < dedup_window:
            hard_reasons.remove(hr)
-            details.append(f"{hr} 近期已触发，{HARD_REASON_DEDUP_MINUTES}分钟内去重")
+            details.append(f"{hr} triggered recently, deduplicated within {HARD_REASON_DEDUP_MINUTES} minutes")

    hard_trigger = bool(hard_reasons)
    if profile.get("dust_mode") and not hard_trigger and soft_score < soft_score_threshold + 1.0:
-        details.append("微型资金/粉尘仓位模式：抬高软触发门槛，避免无意义分析")
+        details.append("Dust-mode portfolio: raising soft-trigger threshold to avoid noise")

    if profile.get("dust_mode") and not profile.get("new_entries_allowed") and any(
        r in {"new-leader", "candidate-leader-init"} for r in soft_reasons
    ):
-        details.append("当前可用资金低于可执行阈值，新候选币仅做观察，不单独触发深度分析")
+        details.append("Available capital below executable threshold; new candidates are observation-only")
        soft_score = max(0.0, soft_score - 0.75)

    should_analyze = hard_trigger or soft_score >= soft_score_threshold

    if cooldown_active and not hard_trigger and should_analyze:
        should_analyze = False
-        details.append(f"处于 {cooldown_minutes} 分钟冷却窗口，软触发先记录不升级")
+        details.append(f"In {cooldown_minutes} minute cooldown window; soft trigger logged but not escalated")

    if cooldown_active and not hard_trigger and reasons and soft_score < soft_score_threshold:
-        details.append(f"处于 {cooldown_minutes} 分钟冷却窗口，且软信号强度不足 ({soft_score:.2f} < {soft_score_threshold:.2f})")
+        details.append(f"In {cooldown_minutes} minute cooldown window with insufficient soft signal ({soft_score:.2f} < {soft_score_threshold:.2f})")

    status = "deep_analysis_required" if should_analyze else "stable"

    compact_lines = [
-        f"状态: {status}",
-        f"组合净值: ${snapshot['portfolio_value_usdt']:.4f} | 可用USDT: ${snapshot['free_usdt']:.4f}",
-        f"本地时段: {snapshot['session']} | 时区: {snapshot['timezone']}",
-        f"BTC/ETH: {market.get('btc_regime')} ({market.get('btc_24h_pct')}%), {market.get('eth_regime')} ({market.get('eth_24h_pct')}%) | 波动分数 {market.get('volatility_score')}",
-        f"门控画像: capital={profile['capital_band']}, session={profile['session_mode']}, volatility={profile['volatility_mode']}, dust={profile['dust_mode']}",
-        f"阈值: price={price_trigger:.1%}, pnl={pnl_trigger:.1%}, portfolio={portfolio_trigger:.1%}, candidate={candidate_ratio_trigger:.2f}, cooldown={effective_cooldown}m({cooldown_minutes}m基础), force={force_minutes}m",
-        f"软信号分: {soft_score:.2f} / {soft_score_threshold:.2f}",
-        f"信号变化度: {signal_delta:.1f}",
+        f"Status: {status}",
+        f"Portfolio: ${snapshot['portfolio_value_usdt']:.4f} | Free USDT: ${snapshot['free_usdt']:.4f}",
+        f"Session: {snapshot['session']} | TZ: {snapshot['timezone']}",
+        f"BTC/ETH: {market.get('btc_regime')} ({market.get('btc_24h_pct')}%), {market.get('eth_regime')} ({market.get('eth_24h_pct')}%) | Volatility score {market.get('volatility_score')}",
+        f"Profile: capital={profile['capital_band']}, session={profile['session_mode']}, volatility={profile['volatility_mode']}, dust={profile['dust_mode']}",
+        f"Thresholds: price={price_trigger:.1%}, pnl={pnl_trigger:.1%}, portfolio={portfolio_trigger:.1%}, candidate={candidate_ratio_trigger:.2f}, cooldown={effective_cooldown}m({cooldown_minutes}m base), force={force_minutes}m",
+        f"Soft score: {soft_score:.2f} / {soft_score_threshold:.2f}",
+        f"Signal delta: {signal_delta:.1f}",
    ]
    if snapshot["positions"]:
-        compact_lines.append("持仓:")
+        compact_lines.append("Positions:")
        for pos in snapshot["positions"][:4]:
            pnl = pos.get("pnl_pct")
            pnl_text = f"{pnl:+.1%}" if pnl is not None else "n/a"
@@ -269,9 +269,9 @@ def analyze_trigger(snapshot: dict, state: dict):
                f"- {pos['symbol']}: qty={pos['quantity']}, px={pos.get('last_price')}, pnl={pnl_text}, value=${pos.get('market_value_usdt')}"
            )
    else:
-        compact_lines.append("持仓: 当前无现货仓位")
+        compact_lines.append("Positions: no spot positions currently")
    if snapshot["top_candidates"]:
-        compact_lines.append("候选榜:")
+        compact_lines.append("Candidates:")
        for cand in snapshot["top_candidates"]:
            compact_lines.append(
                f"- {cand['symbol']}: score={cand['score']}, 24h={cand['change_24h_pct']}%, vol=${cand['volume_24h']}"
@@ -279,13 +279,13 @@ def analyze_trigger(snapshot: dict, state: dict):
    layers = snapshot.get("top_candidates_layers", {})
    for band, band_cands in layers.items():
        if band_cands:
-            compact_lines.append(f"{band} 层:")
+            compact_lines.append(f"{band} tier:")
            for cand in band_cands:
                compact_lines.append(
                    f"- {cand['symbol']}: score={cand['score']}, 24h={cand['change_24h_pct']}%, vol=${cand['volume_24h']}"
                )
    if details:
-        compact_lines.append("触发说明:")
+        compact_lines.append("Trigger notes:")
        for item in details:
            compact_lines.append(f"- {item}")

--- a/src/coinhunter/smart_executor.py
+++ b/src/coinhunter/smart_executor.py
@@ -11,8 +11,6 @@ from __future__ import annotations
 import sys
 from importlib import import_module

-from .services.smart_executor_service import run as _run_service
-
 _EXPORT_MAP = {
    "PATHS": (".runtime", "get_runtime_paths"),
    "ENV_FILE": (".runtime", "get_runtime_paths"),
@@ -39,6 +37,7 @@ _EXPORT_MAP = {
    "save_executions": (".services.execution_state", "save_executions"),
    "load_positions": (".services.portfolio_service", "load_positions"),
    "save_positions": (".services.portfolio_service", "save_positions"),
+    "update_positions": (".services.portfolio_service", "update_positions"),
    "upsert_position": (".services.portfolio_service", "upsert_position"),
    "reconcile_positions_with_exchange": (".services.portfolio_service", "reconcile_positions_with_exchange"),
    "get_exchange": (".services.exchange_service", "get_exchange"),
@@ -58,6 +57,9 @@ _EXPORT_MAP = {
    "action_rebalance": (".services.trade_execution", "action_rebalance"),
    "command_status": (".services.trade_execution", "command_status"),
    "command_balances": (".services.trade_execution", "command_balances"),
+    "command_orders": (".services.trade_execution", "command_orders"),
+    "command_order_status": (".services.trade_execution", "command_order_status"),
+    "command_cancel": (".services.trade_execution", "command_cancel"),
 }

 __all__ = sorted(set(_EXPORT_MAP) | {"ENV_FILE", "PATHS", "load_env", "main"})
@@ -93,6 +95,7 @@ def load_env():


 def main(argv=None):
+    from .services.smart_executor_service import run as _run_service
    return _run_service(sys.argv[1:] if argv is None else argv)


--- a/tests/init.py
+++ b/tests/init.py
--- a/tests/test_check_api.py
+++ b/tests/test_check_api.py
@@ -0,0 +1,70 @@
+"""Tests for check_api command."""
+
+import json
+from unittest.mock import MagicMock, patch
+
+from coinhunter.commands import check_api
+
+
+class TestMain:
+    def test_missing_api_key(self, monkeypatch, capsys):
+        monkeypatch.setenv("BINANCE_API_KEY", "")
+        monkeypatch.setenv("BINANCE_API_SECRET", "secret")
+        rc = check_api.main()
+        assert rc == 1
+        out = json.loads(capsys.readouterr().out)
+        assert out["ok"] is False
+        assert "BINANCE_API_KEY" in out["error"]
+
+    def test_missing_api_secret(self, monkeypatch, capsys):
+        monkeypatch.setenv("BINANCE_API_KEY", "key")
+        monkeypatch.setenv("BINANCE_API_SECRET", "")
+        rc = check_api.main()
+        assert rc == 1
+        out = json.loads(capsys.readouterr().out)
+        assert out["ok"] is False
+        assert "BINANCE_API_SECRET" in out["error"]
+
+    def test_placholder_api_key(self, monkeypatch, capsys):
+        monkeypatch.setenv("BINANCE_API_KEY", "your_api_key")
+        monkeypatch.setenv("BINANCE_API_SECRET", "secret")
+        rc = check_api.main()
+        assert rc == 1
+
+    def test_balance_fetch_failure(self, monkeypatch, capsys):
+        monkeypatch.setenv("BINANCE_API_KEY", "key")
+        monkeypatch.setenv("BINANCE_API_SECRET", "secret")
+        mock_ex = MagicMock()
+        mock_ex.fetch_balance.side_effect = Exception("Network error")
+        with patch("coinhunter.commands.check_api.ccxt.binance", return_value=mock_ex):
+            rc = check_api.main()
+        assert rc == 1
+        out = json.loads(capsys.readouterr().out)
+        assert "Failed to connect" in out["error"]
+
+    def test_success_with_spot_trading(self, monkeypatch, capsys):
+        monkeypatch.setenv("BINANCE_API_KEY", "key")
+        monkeypatch.setenv("BINANCE_API_SECRET", "secret")
+        mock_ex = MagicMock()
+        mock_ex.fetch_balance.return_value = {"USDT": 100.0}
+        mock_ex.sapi_get_account_api_restrictions.return_value = {"enableSpotTrading": True}
+        with patch("coinhunter.commands.check_api.ccxt.binance", return_value=mock_ex):
+            rc = check_api.main()
+        assert rc == 0
+        out = json.loads(capsys.readouterr().out)
+        assert out["ok"] is True
+        assert out["read_permission"] is True
+        assert out["spot_trading_enabled"] is True
+
+    def test_success_restrictions_query_fails(self, monkeypatch, capsys):
+        monkeypatch.setenv("BINANCE_API_KEY", "key")
+        monkeypatch.setenv("BINANCE_API_SECRET", "secret")
+        mock_ex = MagicMock()
+        mock_ex.fetch_balance.return_value = {"USDT": 100.0}
+        mock_ex.sapi_get_account_api_restrictions.side_effect = Exception("no permission")
+        with patch("coinhunter.commands.check_api.ccxt.binance", return_value=mock_ex):
+            rc = check_api.main()
+        assert rc == 0
+        out = json.loads(capsys.readouterr().out)
+        assert out["spot_trading_enabled"] is None
+        assert "may be null" in out["note"]
--- a/tests/test_cli.py
+++ b/tests/test_cli.py
@@ -0,0 +1,58 @@
+"""Tests for CLI routing and parser behavior."""
+
+import pytest
+
+from coinhunter.cli import ALIASES, MODULE_MAP, build_parser, run_python_module
+
+
+class TestAliases:
+    def test_all_aliases_resolve_to_canonical(self):
+        for alias, canonical in ALIASES.items():
+            assert canonical in MODULE_MAP, f"alias {alias!r} points to missing canonical {canonical!r}"
+
+    def test_no_alias_is_itself_an_alias_loop(self):
+        for alias in ALIASES:
+            assert alias not in ALIASES.values() or alias in MODULE_MAP
+
+
+class TestModuleMap:
+    def test_all_modules_exist(self):
+        import importlib
+
+        for command, module_name in MODULE_MAP.items():
+            full = f"coinhunter.{module_name}"
+            mod = importlib.import_module(full)
+            assert hasattr(mod, "main"), f"{full} missing main()"
+
+
+class TestBuildParser:
+    def test_help_includes_commands(self):
+        parser = build_parser()
+        help_text = parser.format_help()
+        assert "coinhunter diag" in help_text
+        assert "coinhunter exec" in help_text
+
+    def test_parses_command_and_args(self):
+        parser = build_parser()
+        ns = parser.parse_args(["exec", "bal"])
+        assert ns.command == "exec"
+        assert "bal" in ns.args
+
+    def test_version_action_exits(self):
+        parser = build_parser()
+        with pytest.raises(SystemExit) as exc:
+            parser.parse_args(["--version"])
+        assert exc.value.code == 0
+
+
+class TestRunPythonModule:
+    def test_runs_module_main_and_returns_int(self):
+        result = run_python_module("commands.paths", [], "coinhunter paths")
+        assert result == 0
+
+    def test_mutates_sys_argv(self):
+        import sys
+
+        original = sys.argv[:]
+        run_python_module("commands.paths", ["--help"], "coinhunter paths")
+        assert sys.argv == original
--- a/tests/test_exchange_service.py
+++ b/tests/test_exchange_service.py
@@ -0,0 +1,98 @@
+"""Tests for exchange helpers and caching."""
+
+import pytest
+
+from coinhunter.services import exchange_service as es
+
+
+class TestNormSymbol:
+    def test_already_normalized(self):
+        assert es.norm_symbol("BTC/USDT") == "BTC/USDT"
+
+    def test_raw_symbol(self):
+        assert es.norm_symbol("BTCUSDT") == "BTC/USDT"
+
+    def test_lowercase(self):
+        assert es.norm_symbol("btcusdt") == "BTC/USDT"
+
+    def test_dash_separator(self):
+        assert es.norm_symbol("BTC-USDT") == "BTC/USDT"
+
+    def test_underscore_separator(self):
+        assert es.norm_symbol("BTC_USDT") == "BTC/USDT"
+
+    def test_unsupported_symbol(self):
+        with pytest.raises(ValueError):
+            es.norm_symbol("BTC")
+
+
+class TestStorageSymbol:
+    def test_converts_to_flat(self):
+        assert es.storage_symbol("BTC/USDT") == "BTCUSDT"
+
+    def test_raw_input(self):
+        assert es.storage_symbol("ETHUSDT") == "ETHUSDT"
+
+
+class TestFloorToStep:
+    def test_no_step(self):
+        assert es.floor_to_step(1.234, 0) == 1.234
+
+    def test_floors_to_step(self):
+        assert es.floor_to_step(1.234, 0.1) == pytest.approx(1.2)
+
+    def test_exact_multiple(self):
+        assert es.floor_to_step(12, 1) == 12
+
+
+class TestGetExchangeCaching:
+    def test_returns_cached_instance(self, monkeypatch):
+        monkeypatch.setenv("BINANCE_API_KEY", "test_key")
+        monkeypatch.setenv("BINANCE_API_SECRET", "test_secret")
+
+        class FakeEx:
+            def load_markets(self):
+                pass
+
+        # Reset cache and patch ccxt.binance
+        original_cache = es._exchange_cache
+        original_cached_at = es._exchange_cached_at
+        try:
+            es._exchange_cache = None
+            es._exchange_cached_at = None
+            monkeypatch.setattr(es.ccxt, "binance", lambda *a, **kw: FakeEx())
+
+            first = es.get_exchange()
+            second = es.get_exchange()
+            assert first is second
+
+            third = es.get_exchange(force_new=True)
+            assert third is not first
+        finally:
+            es._exchange_cache = original_cache
+            es._exchange_cached_at = original_cached_at
+
+    def test_cache_expires_after_ttl(self, monkeypatch):
+        monkeypatch.setenv("BINANCE_API_KEY", "test_key")
+        monkeypatch.setenv("BINANCE_API_SECRET", "test_secret")
+
+        class FakeEx:
+            def load_markets(self):
+                pass
+
+        original_cache = es._exchange_cache
+        original_cached_at = es._exchange_cached_at
+        try:
+            es._exchange_cache = None
+            es._exchange_cached_at = None
+            monkeypatch.setattr(es.ccxt, "binance", lambda *a, **kw: FakeEx())
+            monkeypatch.setattr(es, "load_env", lambda: None)
+
+            first = es.get_exchange()
+            # Simulate time passing beyond TTL
+            es._exchange_cached_at -= es.CACHE_TTL_SECONDS + 1
+            second = es.get_exchange()
+            assert first is not second
+        finally:
+            es._exchange_cache = original_cache
+            es._exchange_cached_at = original_cached_at
--- a/tests/test_external_gate.py
+++ b/tests/test_external_gate.py
@@ -0,0 +1,203 @@
+"""Tests for external_gate configurable hook."""
+
+import json
+from unittest.mock import MagicMock, patch
+
+from coinhunter.commands import external_gate
+
+
+class TestResolveTriggerCommand:
+    def test_missing_config_means_disabled(self, tmp_path, monkeypatch):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        # no config file exists
+        cmd = external_gate._resolve_trigger_command(paths)
+        assert cmd is None
+
+    def test_uses_explicit_list_from_config(self, tmp_path, monkeypatch):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        paths.config_file.write_text(json.dumps({
+            "external_gate": {"trigger_command": ["my-scheduler", "run", "job-123"]}
+        }), encoding="utf-8")
+        cmd = external_gate._resolve_trigger_command(paths)
+        assert cmd == ["my-scheduler", "run", "job-123"]
+
+    def test_null_means_disabled(self, tmp_path, monkeypatch):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        paths.config_file.write_text(json.dumps({
+            "external_gate": {"trigger_command": None}
+        }), encoding="utf-8")
+        cmd = external_gate._resolve_trigger_command(paths)
+        assert cmd is None
+
+    def test_empty_list_means_disabled(self, tmp_path, monkeypatch):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        paths.config_file.write_text(json.dumps({
+            "external_gate": {"trigger_command": []}
+        }), encoding="utf-8")
+        cmd = external_gate._resolve_trigger_command(paths)
+        assert cmd is None
+
+    def test_string_gets_wrapped(self, tmp_path, monkeypatch):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        paths.config_file.write_text(json.dumps({
+            "external_gate": {"trigger_command": "my-script.sh"}
+        }), encoding="utf-8")
+        cmd = external_gate._resolve_trigger_command(paths)
+        assert cmd == ["my-script.sh"]
+
+    def test_unexpected_type_returns_none_and_warns(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        paths.config_file.write_text(json.dumps({
+            "external_gate": {"trigger_command": 42}
+        }), encoding="utf-8")
+        cmd = external_gate._resolve_trigger_command(paths)
+        assert cmd is None
+
+
+class TestMain:
+    def test_already_running(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.state_dir.mkdir(parents=True, exist_ok=True)
+        # Acquire the lock in this process so main() sees it as busy
+        import fcntl
+        with open(paths.external_gate_lock, "w", encoding="utf-8") as f:
+            fcntl.flock(f.fileno(), fcntl.LOCK_EX)
+            rc = external_gate.main()
+        assert rc == 0
+        out = json.loads(capsys.readouterr().out)
+        assert out["reason"] == "already_running"
+
+    def test_precheck_failure(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        fake_result = MagicMock(returncode=1, stdout="err", stderr="")
+        with patch("coinhunter.commands.external_gate.run_cmd", return_value=fake_result):
+            rc = external_gate.main()
+        assert rc == 1
+        out = json.loads(capsys.readouterr().out)
+        assert out["reason"] == "precheck_failed"
+
+    def test_precheck_parse_error(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        fake_result = MagicMock(returncode=0, stdout="not-json", stderr="")
+        with patch("coinhunter.commands.external_gate.run_cmd", return_value=fake_result):
+            rc = external_gate.main()
+        assert rc == 1
+        out = json.loads(capsys.readouterr().out)
+        assert out["reason"] == "precheck_parse_error"
+
+    def test_precheck_not_ok(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        fake_result = MagicMock(returncode=0, stdout=json.dumps({"ok": False}), stderr="")
+        with patch("coinhunter.commands.external_gate.run_cmd", return_value=fake_result):
+            rc = external_gate.main()
+        assert rc == 1
+        out = json.loads(capsys.readouterr().out)
+        assert out["reason"] == "precheck_not_ok"
+
+    def test_no_trigger(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        fake_result = MagicMock(returncode=0, stdout=json.dumps({"ok": True, "should_analyze": False}), stderr="")
+        with patch("coinhunter.commands.external_gate.run_cmd", return_value=fake_result):
+            rc = external_gate.main()
+        assert rc == 0
+        out = json.loads(capsys.readouterr().out)
+        assert out["reason"] == "no_trigger"
+
+    def test_already_queued(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        fake_result = MagicMock(
+            returncode=0,
+            stdout=json.dumps({"ok": True, "should_analyze": True, "run_requested": True, "run_requested_at": "2024-01-01T00:00:00Z"}),
+            stderr="",
+        )
+        with patch("coinhunter.commands.external_gate.run_cmd", return_value=fake_result):
+            rc = external_gate.main()
+        assert rc == 0
+        out = json.loads(capsys.readouterr().out)
+        assert out["reason"] == "already_queued"
+
+    def test_trigger_disabled(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        paths.config_file.write_text(json.dumps({"external_gate": {"trigger_command": None}}), encoding="utf-8")
+        # First call is precheck, second is mark-run-requested
+        responses = [
+            MagicMock(returncode=0, stdout=json.dumps({"ok": True, "should_analyze": True}), stderr=""),
+            MagicMock(returncode=0, stdout=json.dumps({"ok": True}), stderr=""),
+        ]
+        with patch("coinhunter.commands.external_gate.run_cmd", side_effect=responses):
+            rc = external_gate.main()
+        assert rc == 0
+        out = json.loads(capsys.readouterr().out)
+        assert out["reason"] == "trigger_disabled"
+
+    def test_trigger_success(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        paths.config_file.write_text(json.dumps({"external_gate": {"trigger_command": ["echo", "ok"]}}), encoding="utf-8")
+        responses = [
+            MagicMock(returncode=0, stdout=json.dumps({"ok": True, "should_analyze": True, "reasons": ["price-move"]}), stderr=""),
+            MagicMock(returncode=0, stdout=json.dumps({"ok": True}), stderr=""),
+            MagicMock(returncode=0, stdout="triggered", stderr=""),
+        ]
+        with patch("coinhunter.commands.external_gate.run_cmd", side_effect=responses):
+            rc = external_gate.main()
+        assert rc == 0
+        out = json.loads(capsys.readouterr().out)
+        assert out["triggered"] is True
+        assert out["reason"] == "price-move"
+
+    def test_trigger_command_failure(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        paths.config_file.write_text(json.dumps({"external_gate": {"trigger_command": ["false"]}}), encoding="utf-8")
+        responses = [
+            MagicMock(returncode=0, stdout=json.dumps({"ok": True, "should_analyze": True}), stderr=""),
+            MagicMock(returncode=0, stdout=json.dumps({"ok": True}), stderr=""),
+            MagicMock(returncode=1, stdout="", stderr="fail"),
+        ]
+        with patch("coinhunter.commands.external_gate.run_cmd", side_effect=responses):
+            rc = external_gate.main()
+        assert rc == 1
+        out = json.loads(capsys.readouterr().out)
+        assert out["reason"] == "trigger_failed"
+
+    def test_mark_run_requested_failure(self, tmp_path, monkeypatch, capsys):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = external_gate._paths()
+        paths.root.mkdir(parents=True, exist_ok=True)
+        paths.config_file.write_text(json.dumps({"external_gate": {"trigger_command": ["echo", "ok"]}}), encoding="utf-8")
+        responses = [
+            MagicMock(returncode=0, stdout=json.dumps({"ok": True, "should_analyze": True}), stderr=""),
+            MagicMock(returncode=1, stdout="err", stderr=""),
+        ]
+        with patch("coinhunter.commands.external_gate.run_cmd", side_effect=responses):
+            rc = external_gate.main()
+        assert rc == 1
+        out = json.loads(capsys.readouterr().out)
+        assert out["reason"] == "mark_failed"
--- a/tests/test_review_service.py
+++ b/tests/test_review_service.py
@@ -0,0 +1,194 @@
+"""Tests for review_service."""
+
+import json
+from pathlib import Path
+from unittest.mock import patch
+
+from coinhunter.services import review_service as rs
+
+
+class TestAnalyzeTrade:
+    def test_buy_good_when_price_up(self):
+        ex = None
+        trade = {"symbol": "BTC/USDT", "price": 50000.0, "action": "BUY"}
+        with patch.object(rs, "fetch_current_price", return_value=52000.0):
+            result = rs.analyze_trade(trade, ex)
+        assert result["action"] == "BUY"
+        assert result["pnl_estimate_pct"] == 4.0
+        assert result["outcome_assessment"] == "good"
+
+    def test_buy_bad_when_price_down(self):
+        trade = {"symbol": "BTC/USDT", "price": 50000.0, "action": "BUY"}
+        with patch.object(rs, "fetch_current_price", return_value=48000.0):
+            result = rs.analyze_trade(trade, None)
+        assert result["pnl_estimate_pct"] == -4.0
+        assert result["outcome_assessment"] == "bad"
+
+    def test_sell_all_missed_when_price_up(self):
+        trade = {"symbol": "BTC/USDT", "price": 50000.0, "action": "SELL_ALL"}
+        with patch.object(rs, "fetch_current_price", return_value=53000.0):
+            result = rs.analyze_trade(trade, None)
+        assert result["pnl_estimate_pct"] == -6.0
+        assert result["outcome_assessment"] == "missed"
+
+    def test_neutral_when_small_move(self):
+        trade = {"symbol": "BTC/USDT", "price": 50000.0, "action": "BUY"}
+        with patch.object(rs, "fetch_current_price", return_value=50100.0):
+            result = rs.analyze_trade(trade, None)
+        assert result["outcome_assessment"] == "neutral"
+
+    def test_none_when_no_price(self):
+        trade = {"symbol": "BTC/USDT", "action": "BUY"}
+        result = rs.analyze_trade(trade, None)
+        assert result["pnl_estimate_pct"] is None
+        assert result["outcome_assessment"] == "neutral"
+
+
+class TestAnalyzeHoldPasses:
+    def test_finds_passed_opportunities_that_rise(self):
+        decisions = [
+            {
+                "timestamp": "2024-01-01T00:00:00Z",
+                "decision": "HOLD",
+                "analysis": {"opportunities_evaluated": [{"symbol": "ETH/USDT", "verdict": "PASS"}]},
+                "market_snapshot": {"ETHUSDT": {"lastPrice": 100.0}},
+            }
+        ]
+        with patch.object(rs, "fetch_current_price", return_value=110.0):
+            result = rs.analyze_hold_passes(decisions, None)
+        assert len(result) == 1
+        assert result[0]["symbol"] == "ETH/USDT"
+        assert result[0]["change_pct"] == 10.0
+
+    def test_ignores_non_pass_verdicts(self):
+        decisions = [
+            {
+                "decision": "HOLD",
+                "analysis": {"opportunities_evaluated": [{"symbol": "ETH/USDT", "verdict": "BUY"}]},
+                "market_snapshot": {"ETHUSDT": {"lastPrice": 100.0}},
+            }
+        ]
+        with patch.object(rs, "fetch_current_price", return_value=110.0):
+            result = rs.analyze_hold_passes(decisions, None)
+        assert result == []
+
+    def test_ignores_small_moves(self):
+        decisions = [
+            {
+                "decision": "HOLD",
+                "analysis": {"opportunities_evaluated": [{"symbol": "ETH/USDT", "verdict": "PASS"}]},
+                "market_snapshot": {"ETHUSDT": {"lastPrice": 100.0}},
+            }
+        ]
+        with patch.object(rs, "fetch_current_price", return_value=102.0):
+            result = rs.analyze_hold_passes(decisions, None)
+        assert result == []
+
+
+class TestAnalyzeCashMisses:
+    def test_finds_cash_sit_misses(self):
+        decisions = [
+            {
+                "timestamp": "2024-01-01T00:00:00Z",
+                "balances": {"USDT": 100.0},
+                "market_snapshot": {"BTCUSDT": {"lastPrice": 50000.0}},
+            }
+        ]
+        with patch.object(rs, "fetch_current_price", return_value=54000.0):
+            result = rs.analyze_cash_misses(decisions, None)
+        assert len(result) == 1
+        assert result[0]["symbol"] == "BTCUSDT"
+        assert result[0]["change_pct"] == 8.0
+
+    def test_requires_mostly_usdt(self):
+        decisions = [
+            {
+                "balances": {"USDT": 10.0, "BTC": 90.0},
+                "market_snapshot": {"ETHUSDT": {"lastPrice": 100.0}},
+            }
+        ]
+        with patch.object(rs, "fetch_current_price", return_value=200.0):
+            result = rs.analyze_cash_misses(decisions, None)
+        assert result == []
+
+    def test_dedupes_by_best_change(self):
+        decisions = [
+            {
+                "timestamp": "2024-01-01T00:00:00Z",
+                "balances": {"USDT": 100.0},
+                "market_snapshot": {"BTCUSDT": {"lastPrice": 50000.0}},
+            },
+            {
+                "timestamp": "2024-01-01T01:00:00Z",
+                "balances": {"USDT": 100.0},
+                "market_snapshot": {"BTCUSDT": {"lastPrice": 52000.0}},
+            },
+        ]
+        with patch.object(rs, "fetch_current_price", return_value=56000.0):
+            result = rs.analyze_cash_misses(decisions, None)
+        assert len(result) == 1
+        # best change is from 50000 -> 56000 = 12%
+        assert result[0]["change_pct"] == 12.0
+
+
+class TestGenerateReview:
+    def test_empty_period(self, monkeypatch):
+        monkeypatch.setattr(rs, "get_logs_last_n_hours", lambda log_type, hours: [])
+        review = rs.generate_review(1)
+        assert review["total_decisions"] == 0
+        assert review["total_trades"] == 0
+        assert any("No decisions or trades" in i for i in review["insights"])
+
+    def test_with_trades_and_decisions(self, monkeypatch):
+        trades = [
+            {"symbol": "BTC/USDT", "price": 50000.0, "action": "BUY", "timestamp": "2024-01-01T00:00:00Z"}
+        ]
+        decisions = [
+            {
+                "timestamp": "2024-01-01T00:00:00Z",
+                "decision": "HOLD",
+                "analysis": {"opportunities_evaluated": [{"symbol": "ETH/USDT", "verdict": "PASS"}]},
+                "market_snapshot": {"ETHUSDT": {"lastPrice": 100.0}},
+            }
+        ]
+        errors = [{"message": "oops"}]
+
+        def _logs(log_type, hours):
+            return {"decisions": decisions, "trades": trades, "errors": errors}.get(log_type, [])
+
+        monkeypatch.setattr(rs, "get_logs_last_n_hours", _logs)
+        monkeypatch.setattr(rs, "fetch_current_price", lambda ex, sym: 110.0 if "ETH" in sym else 52000.0)
+        review = rs.generate_review(1)
+        assert review["total_trades"] == 1
+        assert review["total_decisions"] == 1
+        assert review["stats"]["good_decisions"] == 1
+        assert review["stats"]["missed_hold_passes"] == 1
+        assert any("execution/system errors" in i for i in review["insights"])
+
+
+class TestSaveReview:
+    def test_writes_file(self, tmp_path, monkeypatch):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        review = {"review_timestamp": "2024-01-01T00:00:00+08:00", "review_period_hours": 1}
+        path = rs.save_review(review)
+        saved = json.loads(Path(path).read_text(encoding="utf-8"))
+        assert saved["review_period_hours"] == 1
+
+
+class TestPrintReview:
+    def test_outputs_report(self, capsys):
+        review = {
+            "review_timestamp": "2024-01-01T00:00:00+08:00",
+            "review_period_hours": 1,
+            "total_decisions": 2,
+            "total_trades": 1,
+            "total_errors": 0,
+            "stats": {"good_decisions": 1, "neutral_decisions": 0, "bad_decisions": 0, "missed_opportunities": 0},
+            "insights": ["Insight A"],
+            "recommendations": ["Rec B"],
+        }
+        rs.print_review(review)
+        captured = capsys.readouterr()
+        assert "Coin Hunter Review Report" in captured.out
+        assert "Insight A" in captured.out
+        assert "Rec B" in captured.out
--- a/tests/test_runtime.py
+++ b/tests/test_runtime.py
@@ -0,0 +1,63 @@
+"""Tests for runtime path resolution."""
+
+from pathlib import Path
+
+import pytest
+
+from coinhunter.runtime import RuntimePaths, ensure_runtime_dirs, get_runtime_paths, mask_secret
+
+
+class TestGetRuntimePaths:
+    def test_defaults_point_to_home_dot_coinhunter(self):
+        paths = get_runtime_paths()
+        assert paths.root == Path.home() / ".coinhunter"
+
+    def test_respects_coinhunter_home_env(self, monkeypatch):
+        monkeypatch.setenv("COINHUNTER_HOME", "~/custom_ch")
+        paths = get_runtime_paths()
+        assert paths.root == Path.home() / "custom_ch"
+
+    def test_respects_hermes_home_env(self, monkeypatch):
+        monkeypatch.setenv("HERMES_HOME", "~/custom_hermes")
+        paths = get_runtime_paths()
+        assert paths.hermes_home == Path.home() / "custom_hermes"
+        assert paths.env_file == Path.home() / "custom_hermes" / ".env"
+
+    def test_returns_frozen_dataclass(self):
+        paths = get_runtime_paths()
+        assert isinstance(paths, RuntimePaths)
+        with pytest.raises(AttributeError):
+            paths.root = Path("/tmp")
+
+    def test_as_dict_returns_strings(self):
+        paths = get_runtime_paths()
+        d = paths.as_dict()
+        assert isinstance(d, dict)
+        assert all(isinstance(v, str) for v in d.values())
+        assert "root" in d
+
+
+class TestEnsureRuntimeDirs:
+    def test_creates_directories(self, tmp_path, monkeypatch):
+        monkeypatch.setenv("COINHUNTER_HOME", str(tmp_path))
+        paths = get_runtime_paths()
+        returned = ensure_runtime_dirs(paths)
+        assert returned.root.exists()
+        assert returned.state_dir.exists()
+        assert returned.logs_dir.exists()
+        assert returned.cache_dir.exists()
+        assert returned.reviews_dir.exists()
+
+
+class TestMaskSecret:
+    def test_empty_string(self):
+        assert mask_secret("") == ""
+
+    def test_short_value(self):
+        assert mask_secret("ab") == "**"
+
+    def test_masks_all_but_tail(self):
+        assert mask_secret("supersecret", tail=4) == "*******cret"
+
+    def test_none_returns_empty(self):
+        assert mask_secret(None) == ""
--- a/tests/test_smart_executor_service.py
+++ b/tests/test_smart_executor_service.py
@@ -0,0 +1,98 @@
+"""Tests for smart executor deduplication and routing."""
+
+
+from coinhunter.services import exchange_service
+from coinhunter.services import smart_executor_service as ses
+
+
+class FakeArgs:
+    def __init__(self, command="buy", symbol="BTCUSDT", amount_usdt=50, dry_run=False,
+                 decision_id=None, analysis=None, reasoning=None, order_id=None,
+                 from_symbol=None, to_symbol=None):
+        self.command = command
+        self.symbol = symbol
+        self.amount_usdt = amount_usdt
+        self.dry_run = dry_run
+        self.decision_id = decision_id
+        self.analysis = analysis
+        self.reasoning = reasoning
+        self.order_id = order_id
+        self.from_symbol = from_symbol
+        self.to_symbol = to_symbol
+
+
+class TestDeduplication:
+    def test_skips_duplicate_mutating_action(self, monkeypatch, capsys):
+        monkeypatch.setattr(ses, "parse_cli_args", lambda argv: (FakeArgs(command="buy"), ["BTCUSDT", "50"]))
+        monkeypatch.setattr(ses, "get_execution_state", lambda did: {"status": "success"})
+        get_exchange_calls = []
+        monkeypatch.setattr(exchange_service, "get_exchange", lambda: get_exchange_calls.append(1) or "ex")
+
+        result = ses.run(["buy", "BTCUSDT", "50"])
+        assert result == 0
+        assert get_exchange_calls == []
+        captured = capsys.readouterr()
+        assert "already executed successfully" in captured.err
+
+    def test_allows_duplicate_read_only_action(self, monkeypatch):
+        monkeypatch.setattr(ses, "parse_cli_args", lambda argv: (FakeArgs(command="balances"), ["balances"]))
+        monkeypatch.setattr(ses, "get_execution_state", lambda did: {"status": "success"})
+        monkeypatch.setattr(ses, "command_balances", lambda ex: None)
+        get_exchange_calls = []
+        monkeypatch.setattr(exchange_service, "get_exchange", lambda: get_exchange_calls.append(1) or "ex")
+
+        result = ses.run(["bal"])
+        assert result == 0
+        assert len(get_exchange_calls) == 1
+
+    def test_allows_retry_after_failure(self, monkeypatch):
+        monkeypatch.setattr(ses, "parse_cli_args", lambda argv: (FakeArgs(command="buy"), ["BTCUSDT", "50"]))
+        monkeypatch.setattr(ses, "get_execution_state", lambda did: {"status": "failed"})
+        monkeypatch.setattr(ses, "record_execution_state", lambda did, state: None)
+        monkeypatch.setattr(ses, "build_decision_context", lambda ex, action, tail, did: {})
+        monkeypatch.setattr(ses, "action_buy", lambda *a, **k: {"id": "123"})
+        monkeypatch.setattr(ses, "print_json", lambda d: None)
+        get_exchange_calls = []
+        monkeypatch.setattr(exchange_service, "get_exchange", lambda: get_exchange_calls.append(1) or "ex")
+
+        result = ses.run(["buy", "BTCUSDT", "50"])
+        assert result == 0
+        assert len(get_exchange_calls) == 1
+
+
+class TestReadOnlyRouting:
+    def test_routes_balances(self, monkeypatch):
+        monkeypatch.setattr(ses, "parse_cli_args", lambda argv: (FakeArgs(command="balances"), ["balances"]))
+        monkeypatch.setattr(ses, "get_execution_state", lambda did: None)
+        routed = []
+        monkeypatch.setattr(ses, "command_balances", lambda ex: routed.append("balances"))
+        monkeypatch.setattr(exchange_service, "get_exchange", lambda: "ex")
+
+        result = ses.run(["balances"])
+        assert result == 0
+        assert routed == ["balances"]
+
+    def test_routes_orders(self, monkeypatch):
+        monkeypatch.setattr(ses, "parse_cli_args", lambda argv: (FakeArgs(command="orders"), ["orders"]))
+        monkeypatch.setattr(ses, "get_execution_state", lambda did: None)
+        routed = []
+        monkeypatch.setattr(ses, "command_orders", lambda ex: routed.append("orders"))
+        monkeypatch.setattr(exchange_service, "get_exchange", lambda: "ex")
+
+        result = ses.run(["orders"])
+        assert result == 0
+        assert routed == ["orders"]
+
+    def test_routes_order_status(self, monkeypatch):
+        monkeypatch.setattr(ses, "parse_cli_args", lambda argv: (
+            FakeArgs(command="order-status", symbol="BTCUSDT", order_id="123"),
+            ["order-status", "BTCUSDT", "123"]
+        ))
+        monkeypatch.setattr(ses, "get_execution_state", lambda did: None)
+        routed = []
+        monkeypatch.setattr(ses, "command_order_status", lambda ex, sym, oid: routed.append((sym, oid)))
+        monkeypatch.setattr(exchange_service, "get_exchange", lambda: "ex")
+
+        result = ses.run(["order-status", "BTCUSDT", "123"])
+        assert result == 0
+        assert routed == [("BTCUSDT", "123")]
--- a/tests/test_state_manager.py
+++ b/tests/test_state_manager.py
@@ -0,0 +1,99 @@
+"""Tests for state_manager precheck utilities."""
+
+from datetime import timedelta, timezone
+
+from coinhunter.services.state_manager import (
+    clear_run_request_fields,
+    sanitize_state_for_stale_triggers,
+    update_state_after_observation,
+)
+from coinhunter.services.time_utils import utc_iso, utc_now
+
+
+class TestClearRunRequestFields:
+    def test_removes_run_fields(self):
+        state = {"run_requested_at": utc_iso(), "run_request_note": "test"}
+        clear_run_request_fields(state)
+        assert "run_requested_at" not in state
+        assert "run_request_note" not in state
+
+
+class TestSanitizeStateForStaleTriggers:
+    def test_no_changes_when_clean(self):
+        state = {"pending_trigger": False}
+        result = sanitize_state_for_stale_triggers(state)
+        assert result["pending_trigger"] is False
+        assert result["_stale_recovery_notes"] == []
+
+    def test_clears_completed_run_request(self):
+        state = {
+            "pending_trigger": True,
+            "run_requested_at": utc_iso(),
+            "last_deep_analysis_at": utc_iso(),
+        }
+        result = sanitize_state_for_stale_triggers(state)
+        assert result["pending_trigger"] is False
+        assert "run_requested_at" not in result
+        assert any("completed run_requested" in note for note in result["_stale_recovery_notes"])
+
+    def test_clears_stale_run_request(self):
+        old = (utc_now() - timedelta(minutes=60)).replace(tzinfo=timezone.utc).isoformat()
+        state = {
+            "pending_trigger": False,
+            "run_requested_at": old,
+        }
+        result = sanitize_state_for_stale_triggers(state)
+        assert "run_requested_at" not in result
+        assert any("stale run_requested" in note for note in result["_stale_recovery_notes"])
+
+    def test_recovers_stale_pending_trigger(self):
+        old = (utc_now() - timedelta(minutes=60)).replace(tzinfo=timezone.utc).isoformat()
+        state = {
+            "pending_trigger": True,
+            "last_triggered_at": old,
+        }
+        result = sanitize_state_for_stale_triggers(state)
+        assert result["pending_trigger"] is False
+        assert any("stale pending_trigger" in note for note in result["_stale_recovery_notes"])
+
+
+class TestUpdateStateAfterObservation:
+    def test_updates_last_observed_fields(self):
+        state = {}
+        snapshot = {
+            "generated_at": "2024-01-01T00:00:00Z",
+            "snapshot_hash": "abc",
+            "positions_hash": "pos123",
+            "candidates_hash": "can456",
+            "portfolio_value_usdt": 100.0,
+            "market_regime": "neutral",
+            "positions": [],
+            "top_candidates": [],
+        }
+        analysis = {"should_analyze": False, "details": [], "adaptive_profile": {}}
+        result = update_state_after_observation(state, snapshot, analysis)
+        assert result["last_observed_at"] == snapshot["generated_at"]
+        assert result["last_snapshot_hash"] == "abc"
+
+    def test_sets_pending_trigger_when_should_analyze(self):
+        state = {}
+        snapshot = {
+            "generated_at": "2024-01-01T00:00:00Z",
+            "snapshot_hash": "abc",
+            "positions_hash": "pos123",
+            "candidates_hash": "can456",
+            "portfolio_value_usdt": 100.0,
+            "market_regime": "neutral",
+            "positions": [],
+            "top_candidates": [],
+        }
+        analysis = {
+            "should_analyze": True,
+            "details": ["price move"],
+            "adaptive_profile": {},
+            "hard_reasons": ["moon"],
+            "signal_delta": 1.5,
+        }
+        result = update_state_after_observation(state, snapshot, analysis)
+        assert result["pending_trigger"] is True
+        assert result["pending_reasons"] == ["price move"]
--- a/tests/test_trade_execution.py
+++ b/tests/test_trade_execution.py
@@ -0,0 +1,160 @@
+"""Tests for trade execution dry-run paths."""
+
+import pytest
+
+from coinhunter.services import trade_execution as te
+from coinhunter.services.trade_common import set_dry_run
+
+
+class TestMarketSellDryRun:
+    def test_returns_dry_run_order(self, monkeypatch):
+        set_dry_run(True)
+        monkeypatch.setattr(
+            te, "prepare_sell_quantity",
+            lambda ex, sym, qty: ("BTC/USDT", 0.5, 60000.0, 30000.0)
+        )
+        order = te.market_sell(None, "BTCUSDT", 0.5, "dec-123")
+        assert order["id"] == "dry-sell-dec-123"
+        assert order["symbol"] == "BTC/USDT"
+        assert order["amount"] == 0.5
+        assert order["status"] == "closed"
+        set_dry_run(False)
+
+
+class TestMarketBuyDryRun:
+    def test_returns_dry_run_order(self, monkeypatch):
+        set_dry_run(True)
+        monkeypatch.setattr(
+            te, "prepare_buy_quantity",
+            lambda ex, sym, amt: ("ETH/USDT", 1.0, 3000.0, 3000.0)
+        )
+        order = te.market_buy(None, "ETHUSDT", 100, "dec-456")
+        assert order["id"] == "dry-buy-dec-456"
+        assert order["symbol"] == "ETH/USDT"
+        assert order["amount"] == 1.0
+        assert order["status"] == "closed"
+        set_dry_run(False)
+
+
+class TestActionSellAll:
+    def test_raises_when_balance_zero(self, monkeypatch):
+        monkeypatch.setattr(te, "fetch_balances", lambda ex: {"BTC": 0})
+        with pytest.raises(RuntimeError, match="balance is zero"):
+            te.action_sell_all(None, "BTCUSDT", "dec-789", {})
+
+    def test_dry_run_does_not_reconcile(self, monkeypatch):
+        set_dry_run(True)
+        monkeypatch.setattr(te, "fetch_balances", lambda ex: {"BTC": 0.5})
+        monkeypatch.setattr(
+            te, "market_sell",
+            lambda ex, sym, qty, did: {"id": "dry-1", "amount": qty, "price": 60000.0, "cost": 30000.0, "status": "closed"}
+        )
+        monkeypatch.setattr(te, "reconcile_positions_with_exchange", lambda ex, hint=None: ([], {}))
+        monkeypatch.setattr(te, "log_trade", lambda *a, **k: None)
+        monkeypatch.setattr(te, "log_decision", lambda *a, **k: None)
+
+        result = te.action_sell_all(None, "BTCUSDT", "dec-789", {"analysis": "test"})
+        assert result["id"] == "dry-1"
+        set_dry_run(False)
+
+
+class TestActionBuy:
+    def test_raises_when_insufficient_usdt(self, monkeypatch):
+        monkeypatch.setattr(te, "fetch_balances", lambda ex: {"USDT": 10.0})
+        with pytest.raises(RuntimeError, match="Insufficient USDT"):
+            te.action_buy(None, "BTCUSDT", 50, "dec-abc", {})
+
+    def test_dry_run_skips_save(self, monkeypatch):
+        set_dry_run(True)
+        monkeypatch.setattr(te, "fetch_balances", lambda ex: {"USDT": 100.0})
+        monkeypatch.setattr(
+            te, "market_buy",
+            lambda ex, sym, amt, did: {"id": "dry-2", "amount": 0.01, "price": 50000.0, "cost": 500.0, "status": "closed"}
+        )
+        monkeypatch.setattr(te, "load_positions", lambda: [])
+        monkeypatch.setattr(te, "upsert_position", lambda positions, pos: positions.append(pos))
+        monkeypatch.setattr(te, "reconcile_positions_with_exchange", lambda ex, hint=None: ([], {}))
+        monkeypatch.setattr(te, "log_trade", lambda *a, **k: None)
+        monkeypatch.setattr(te, "log_decision", lambda *a, **k: None)
+
+        result = te.action_buy(None, "BTCUSDT", 50, "dec-abc", {})
+        assert result["id"] == "dry-2"
+        set_dry_run(False)
+
+
+class TestCommandBalances:
+    def test_returns_balances(self, monkeypatch, capsys):
+        monkeypatch.setattr(te, "fetch_balances", lambda ex: {"USDT": 100.0, "BTC": 0.5})
+        result = te.command_balances(None)
+        assert result == {"USDT": 100.0, "BTC": 0.5}
+        captured = capsys.readouterr()
+        assert '"USDT": 100.0' in captured.out
+
+
+class TestCommandStatus:
+    def test_returns_snapshot(self, monkeypatch, capsys):
+        monkeypatch.setattr(te, "fetch_balances", lambda ex: {"USDT": 100.0})
+        monkeypatch.setattr(te, "load_positions", lambda: [])
+        monkeypatch.setattr(te, "build_market_snapshot", lambda ex: {"BTC/USDT": 50000.0})
+        result = te.command_status(None)
+        assert result["balances"]["USDT"] == 100.0
+        captured = capsys.readouterr()
+        assert '"balances"' in captured.out
+
+
+class TestCommandOrders:
+    def test_returns_orders(self, monkeypatch, capsys):
+        orders = [{"id": "1", "symbol": "BTC/USDT"}]
+        monkeypatch.setattr(te, "print_json", lambda d: None)
+        ex = type("Ex", (), {"fetch_open_orders": lambda self: orders})()
+        result = te.command_orders(ex)
+        assert result == orders
+
+
+class TestCommandOrderStatus:
+    def test_returns_order(self, monkeypatch, capsys):
+        order = {"id": "123", "status": "open"}
+        monkeypatch.setattr(te, "print_json", lambda d: None)
+        ex = type("Ex", (), {"fetch_order": lambda self, oid, sym: order})()
+        result = te.command_order_status(ex, "BTCUSDT", "123")
+        assert result == order
+
+
+class TestCommandCancel:
+    def test_dry_run_returns_early(self, monkeypatch):
+        set_dry_run(True)
+        monkeypatch.setattr(te, "log", lambda msg: None)
+        result = te.command_cancel(None, "BTCUSDT", "123")
+        assert result["dry_run"] is True
+        set_dry_run(False)
+
+    def test_cancel_by_order_id(self, monkeypatch):
+        set_dry_run(False)
+        monkeypatch.setattr(te, "print_json", lambda d: None)
+        ex = type("Ex", (), {"cancel_order": lambda self, oid, sym: {"id": oid, "status": "canceled"}})()
+        result = te.command_cancel(ex, "BTCUSDT", "123")
+        assert result["status"] == "canceled"
+
+    def test_cancel_latest_when_no_order_id(self, monkeypatch):
+        set_dry_run(False)
+        monkeypatch.setattr(te, "print_json", lambda d: None)
+        ex = type("Ex", (), {
+            "fetch_open_orders": lambda self, sym: [{"id": "999"}, {"id": "888"}],
+            "cancel_order": lambda self, oid, sym: {"id": oid, "status": "canceled"},
+        })()
+        result = te.command_cancel(ex, "BTCUSDT", None)
+        assert result["id"] == "888"
+
+    def test_cancel_raises_when_no_open_orders(self, monkeypatch):
+        set_dry_run(False)
+        ex = type("Ex", (), {"fetch_open_orders": lambda self, sym: []})()
+        with pytest.raises(RuntimeError, match="No open orders"):
+            te.command_cancel(ex, "BTCUSDT", None)
+
+
+class TestPrintJson:
+    def test_outputs_sorted_json(self, capsys):
+        te.print_json({"b": 2, "a": 1})
+        captured = capsys.readouterr()
+        assert '"a": 1' in captured.out
+        assert '"b": 2' in captured.out