diff --git a/code_stats.py b/code_stats.py
new file mode 100644
index 0000000000..c6567a111b
--- /dev/null
+++ b/code_stats.py
@@ -0,0 +1,217 @@
+#!/usr/bin/env python3
+"""代码统计分析工具 - 按目录/类型分组，区分代码行/注释行/空行，ASCII柱状图"""
+
+import os
+import sys
+from pathlib import Path
+from collections import defaultdict
+
+# 文件类型 → (单行注释, 多行开始, 多行结束)
+LANG_MAP = {
+    ".py":   ("#", '"""', '"""'),
+    ".js":   ("//", "/*", "*/"),
+    ".ts":   ("//", "/*", "*/"),
+    ".tsx":  ("//", "/*", "*/"),
+    ".jsx":  ("//", "/*", "*/"),
+    ".java": ("//", "/*", "*/"),
+    ".c":    ("//", "/*", "*/"),
+    ".cpp":  ("//", "/*", "*/"),
+    ".h":    ("//", "/*", "*/"),
+    ".go":   ("//", "/*", "*/"),
+    ".rs":   ("//", "/*", "*/"),
+    ".rb":   ("#", "=begin", "=end"),
+    ".sh":   ("#", None, None),
+    ".bash": ("#", None, None),
+    ".yml":  ("#", None, None),
+    ".yaml": ("#", None, None),
+    ".toml": ("#", None, None),
+    ".sql":  ("--", "/*", "*/"),
+    ".html": (None, "<!--", "-->"),
+    ".css":  (None, "/*", "*/"),
+    ".vue":  ("//", "/*", "*/"),
+    ".swift": ("//", "/*", "*/"),
+    ".kt":   ("//", "/*", "*/"),
+    ".lua":  ("--", "--[[", "]]"),
+    ".r":    ("#", None, None),
+    ".php":  ("//", "/*", "*/"),
+}
+
+SKIP_DIRS = {
+    ".git", ".svn", ".hg", "node_modules", "__pycache__",
+    ".venv", "venv", "env", ".env", ".tox", "dist", "build",
+    ".mypy_cache", ".pytest_cache", ".eggs", "target", "vendor",
+    ".next", ".nuxt", "coverage",
+}
+
+
+def should_skip(path: Path) -> bool:
+    return any(part in SKIP_DIRS for part in path.parts)
+
+
+def analyze_file(filepath: Path) -> dict:
+    ext = filepath.suffix.lower()
+    if ext not in LANG_MAP:
+        return None
+
+    try:
+        with open(filepath, "r", encoding="utf-8", errors="ignore") as f:
+            lines = f.readlines()
+    except (OSError, PermissionError):
+        return None
+
+    line_comment, block_start, block_end = LANG_MAP[ext]
+    total = len(lines)
+    blank = 0
+    comment = 0
+    in_block = False
+
+    for line in lines:
+        stripped = line.strip()
+        if not stripped:
+            blank += 1
+            continue
+        if in_block:
+            comment += 1
+            if block_end and block_end in stripped:
+                in_block = False
+            continue
+        if block_start and stripped.startswith(block_start):
+            comment += 1
+            if block_end and block_end not in stripped[len(block_start):]:
+                in_block = True
+            continue
+        if line_comment and stripped.startswith(line_comment):
+            comment += 1
+
+    return {"ext": ext, "total": total, "code": total - blank - comment, "comment": comment, "blank": blank}
+
+
+def bar(value, max_val, width=25):
+    if max_val == 0:
+        return ""
+    return "█" * int(width * value / max_val) + "░" * (width - int(width * value / max_val))
+
+
+def stacked_bar(code, comment, blank, total, width=30):
+    """三段式柱状图：█代码 ▓注释 ░空行"""
+    if total == 0:
+        return "░" * width
+    cw = int(width * code / total)
+    mw = int(width * comment / total)
+    bw = width - cw - mw
+    return "█" * cw + "▓" * mw + "░" * bw
+
+
+def fmt(n):
+    """数字格式化"""
+    return f"{n:,}"
+
+
+def print_table(title, rows, headers, col_widths, grand_row=None):
+    """通用表格打印"""
+    print(f"\n📁 {title}")
+    header_line = "".join(h.rjust(w) for h, w in zip(headers, col_widths))
+    print(header_line)
+    print("─" * len(header_line))
+
+    for row in rows:
+        print("".join(str(v).rjust(w) for v, w in zip(row, col_widths)))
+
+    if grand_row:
+        print("─" * len(header_line))
+        print("".join(str(v).rjust(w) for v, w in zip(grand_row, col_widths)))
+
+
+def main():
+    target = Path(sys.argv[1]) if len(sys.argv) > 1 else Path(".")
+    if not target.is_dir():
+        print(f"❌ 不是有效目录: {target}")
+        sys.exit(1)
+
+    # 收集所有文件的分析结果
+    file_results = []
+    for filepath in target.rglob("*"):
+        if not filepath.is_file() or should_skip(filepath):
+            continue
+        result = analyze_file(filepath)
+        if result:
+            rel = filepath.relative_to(target)
+            file_results.append({"path": rel, **result})
+
+    if not file_results:
+        print("未找到可分析的代码文件。")
+        return
+
+    # 汇总
+    grand = {"files": len(file_results), "total": 0, "code": 0, "comment": 0, "blank": 0}
+    for r in file_results:
+        for k in ("total", "code", "comment", "blank"):
+            grand[k] += r[k]
+
+    comment_rate = grand["comment"] / max(grand["code"] + grand["comment"], 1) * 100
+
+    print(f"\n📊 代码统计 — {target.resolve()}")
+    print(f"   {grand['files']} 个文件 | {fmt(grand['total'])} 行 (代码 {fmt(grand['code'])} / 注释 {fmt(grand['comment'])} / 空行 {fmt(grand['blank'])})")
+    print(f"   代码占比 {grand['code']/max(grand['total'],1)*100:.1f}% | 注释率 {comment_rate:.1f}%")
+
+    # ── 按目录分组 ──
+    dir_stats = defaultdict(lambda: {"files": 0, "total": 0, "code": 0, "comment": 0, "blank": 0})
+    for r in file_results:
+        top_dir = r["path"].parts[0] if len(r["path"].parts) > 1 else "[root]"
+        bucket = dir_stats[top_dir]
+        bucket["files"] += 1
+        for k in ("total", "code", "comment", "blank"):
+            bucket[k] += r[k]
+
+    headers = ["目录", "文件数", "代码行", "注释行", "空行", "注释率"]
+    col_widths = [16, 8, 10, 10, 10, 9]
+    rows = []
+    for d, b in sorted(dir_stats.items(), key=lambda x: x[1]["code"], reverse=True):
+        cr = b["comment"] / max(b["code"] + b["comment"], 1) * 100
+        rows.append([d, str(b["files"]), fmt(b["code"]), fmt(b["comment"]), fmt(b["blank"]), f"{cr:.1f}%"])
+
+    print_table("按目录分组", rows, headers, col_widths,
+                grand_row=["合计", str(grand["files"]), fmt(grand["code"]), fmt(grand["comment"]), fmt(grand["blank"]), f"{comment_rate:.1f}%"])
+
+    # 按目录柱状图
+    max_code = max(b["code"] for b in dir_stats.values())
+    print(f"\n📊 目录代码量分布")
+    for d, b in sorted(dir_stats.items(), key=lambda x: x[1]["code"], reverse=True):
+        pct = b["code"] / max(grand["code"], 1) * 100
+        print(f"  {d:<14} {bar(b['code'], max_code, 30)} {pct:5.1f}%")
+
+    # ── 按文件类型 ──
+    ext_stats = defaultdict(lambda: {"files": 0, "total": 0, "code": 0, "comment": 0, "blank": 0})
+    for r in file_results:
+        bucket = ext_stats[r["ext"]]
+        bucket["files"] += 1
+        for k in ("total", "code", "comment", "blank"):
+            bucket[k] += r[k]
+
+    headers2 = ["类型", "文件数", "代码行", "注释行", "空行", "注释率"]
+    col_widths2 = [10, 8, 10, 10, 10, 9]
+    rows2 = []
+    for ext, b in sorted(ext_stats.items(), key=lambda x: x[1]["code"], reverse=True):
+        cr = b["comment"] / max(b["code"] + b["comment"], 1) * 100
+        rows2.append([ext, str(b["files"]), fmt(b["code"]), fmt(b["comment"]), fmt(b["blank"]), f"{cr:.1f}%"])
+
+    print_table("按文件类型", rows2, headers2, col_widths2,
+                grand_row=["合计", str(grand["files"]), fmt(grand["code"]), fmt(grand["comment"]), fmt(grand["blank"]), f"{comment_rate:.1f}%"])
+
+    # 按类型柱状图
+    max_code_ext = max(b["code"] for b in ext_stats.values())
+    print(f"\n📊 类型代码量分布")
+    for ext, b in sorted(ext_stats.items(), key=lambda x: x[1]["code"], reverse=True):
+        pct = b["code"] / max(grand["code"], 1) * 100
+        print(f"  {ext:<8} {bar(b['code'], max_code_ext, 30)} {pct:5.1f}%")
+
+    # ── 综合堆叠柱状图（按目录） ──
+    print(f"\n📊 目录代码结构（█代码 ▓注释 ░空行）")
+    for d, b in sorted(dir_stats.items(), key=lambda x: x[1]["code"], reverse=True):
+        print(f"  {d:<14} {stacked_bar(b['code'], b['comment'], b['blank'], b['total'], 40)}")
+
+    print()
+
+
+if __name__ == "__main__":
+    main()
diff --git a/count_lines.py b/count_lines.py
new file mode 100644
index 0000000000..ec14b707bf
--- /dev/null
+++ b/count_lines.py
@@ -0,0 +1,85 @@
+#!/usr/bin/env python3
+"""统计当前目录下各类型文件的代码行数"""
+
+import os
+from pathlib import Path
+from collections import defaultdict
+
+# 常见代码文件扩展名
+CODE_EXTENSIONS = {
+    ".py", ".js", ".ts", ".tsx", ".jsx", ".java", ".c", ".cpp", ".h", ".hpp",
+    ".go", ".rs", ".rb", ".php", ".swift", ".kt", ".scala", ".sh", ".bash",
+    ".sql", ".html", ".css", ".scss", ".less", ".vue", ".svelte",
+    ".yaml", ".yml", ".json", ".xml", ".toml", ".ini", ".cfg",
+    ".md", ".rst", ".txt",
+}
+
+# 跳过的目录
+SKIP_DIRS = {
+    "node_modules", ".git", "__pycache__", ".venv", "venv", "env",
+    ".tox", ".mypy_cache", ".pytest_cache", "dist", "build", ".eggs",
+    ".idea", ".vscode", "target", "vendor", ".next", ".nuxt",
+}
+
+
+def count_lines(path: Path) -> tuple[int, int, int]:
+    """返回 (总行数, 非空行数, 注释行数)"""
+    total = blank = comment = 0
+    try:
+        with open(path, "r", encoding="utf-8", errors="ignore") as f:
+            for line in f:
+                total += 1
+                stripped = line.strip()
+                if not stripped:
+                    blank += 1
+                elif stripped.startswith("#") or stripped.startswith("//"):
+                    comment += 1
+    except (PermissionError, OSError):
+        pass
+    return total, total - blank, comment
+
+
+def main():
+    base = Path.cwd()
+    stats = defaultdict(lambda: {"files": 0, "total": 0, "code": 0, "comment": 0})
+
+    for root, dirs, files in os.walk(base):
+        # 原地删除跳过目录
+        dirs[:] = [d for d in dirs if d not in SKIP_DIRS and not d.startswith(".")]
+        for fname in files:
+            fpath = Path(root) / fname
+            ext = fpath.suffix.lower()
+            if ext in CODE_EXTENSIONS:
+                total, code, comment = count_lines(fpath)
+                if total > 0:
+                    lang = ext.lstrip(".")
+                    stats[lang]["files"] += 1
+                    stats[lang]["total"] += total
+                    stats[lang]["code"] += code
+                    stats[lang]["comment"] += comment
+
+    if not stats:
+        print("未找到代码文件")
+        return
+
+    # 按代码行数降序排列
+    sorted_stats = sorted(stats.items(), key=lambda x: x[1]["code"], reverse=True)
+
+    # 表头
+    print(f"\n{'语言':<12} {'文件数':>6} {'总行数':>8} {'代码行':>8} {'注释行':>8}")
+    print("-" * 50)
+
+    sum_files = sum_total = sum_code = sum_comment = 0
+    for lang, s in sorted_stats:
+        print(f".{lang:<11} {s['files']:>6} {s['total']:>8} {s['code']:>8} {s['comment']:>8}")
+        sum_files += s["files"]
+        sum_total += s["total"]
+        sum_code += s["code"]
+        sum_comment += s["comment"]
+
+    print("-" * 50)
+    print(f"{'合计':<12} {sum_files:>6} {sum_total:>8} {sum_code:>8} {sum_comment:>8}\n")
+
+
+if __name__ == "__main__":
+    main()
diff --git a/scripts/claude_acp_bridge.py b/scripts/claude_acp_bridge.py
new file mode 100755
index 0000000000..3d1d761ca4
--- /dev/null
+++ b/scripts/claude_acp_bridge.py
@@ -0,0 +1,179 @@
+#!/usr/bin/env python3
+"""ACP JSON-RPC bridge for Claude Code CLI.
+
+Translates Hermes CopilotACPClient's ACP JSON-RPC protocol into
+`claude -p` CLI calls.  This lets Hermes delegate tasks to Claude Code
+without modifying any upstream Hermes source code.
+
+Usage (via HERMES_COPILOT_ACP_COMMAND env var):
+    HERMES_COPILOT_ACP_COMMAND=python3 scripts/claude_acp_bridge.py
+
+Protocol flow:
+    1. Client sends "initialize" → respond with server info
+    2. Client sends "session/new" → respond with sessionId
+    3. Client sends "session/prompt" → run `claude -p`, stream back result
+"""
+
+import json
+import os
+import subprocess
+import sys
+import uuid
+
+
+def _read_request():
+    """Read one JSON-RPC request from stdin (line-delimited)."""
+    for line in sys.stdin:
+        line = line.strip()
+        if not line:
+            continue
+        try:
+            return json.loads(line)
+        except json.JSONDecodeError:
+            continue
+    return None
+
+
+def _write_msg(msg):
+    """Write a JSON-RPC message to stdout."""
+    sys.stdout.write(json.dumps(msg, ensure_ascii=False) + "\n")
+    sys.stdout.flush()
+
+
+def _send_notification(method, params):
+    """Send a JSON-RPC notification (no id)."""
+    _write_msg({"jsonrpc": "2.0", "method": method, "params": params})
+
+
+def _send_response(request_id, result):
+    """Send a JSON-RPC success response."""
+    _write_msg({"jsonrpc": "2.0", "id": request_id, "result": result})
+
+
+def _send_error(request_id, code, message):
+    """Send a JSON-RPC error response."""
+    _write_msg({
+        "jsonrpc": "2.0",
+        "id": request_id,
+        "error": {"code": code, "message": message},
+    })
+
+
+def _handle_initialize(request_id, _params):
+    _send_response(request_id, {
+        "protocolVersion": 1,
+        "serverCapabilities": {},
+        "serverInfo": {
+            "name": "claude-code-bridge",
+            "title": "Claude Code Bridge",
+            "version": "0.1.0",
+        },
+    })
+
+
+def _handle_session_new(request_id, params):
+    session_id = str(uuid.uuid4())
+    cwd = params.get("cwd", os.getcwd())
+    _send_response(request_id, {"sessionId": session_id, "cwd": cwd})
+
+
+def _handle_session_prompt(request_id, params):
+    # Extract prompt text from ACP format
+    prompt_parts = params.get("prompt", [])
+    prompt_text = ""
+    for part in prompt_parts:
+        if isinstance(part, dict) and part.get("type") == "text":
+            prompt_text += part.get("text", "")
+
+    if not prompt_text:
+        _send_error(request_id, -32602, "Empty prompt")
+        return
+
+    # Build claude CLI command
+    cmd = ["claude", "-p", prompt_text, "--output-format", "json", "--dangerously-skip-permissions"]
+
+    # Add model if specified via env var
+    model = os.getenv("CLAUDE_CODE_BRIDGE_MODEL", "").strip()
+    if model:
+        cmd.extend(["--model", model])
+
+    # Use cwd from session if available
+    cwd = params.get("cwd") or os.getcwd()
+
+    try:
+        result = subprocess.run(
+            cmd,
+            capture_output=True,
+            text=True,
+            timeout=900,
+            cwd=cwd if os.path.isdir(cwd) else None,
+        )
+        if result.returncode != 0:
+            error_msg = result.stderr.strip() or f"claude exited with code {result.returncode}"
+            _send_notification("session/update", {
+                "update": {
+                    "sessionUpdate": "agent_message_chunk",
+                    "content": {"text": f"Error from Claude Code: {error_msg}"},
+                }
+            })
+            _send_response(request_id, {"status": "error"})
+            return
+
+        response_data = json.loads(result.stdout)
+        response_text = response_data.get("result", "")
+
+    except subprocess.TimeoutExpired:
+        response_text = "Error: Claude Code timed out (900s limit)"
+    except json.JSONDecodeError:
+        response_text = result.stdout if result.stdout else "Error: Could not parse Claude Code output"
+    except FileNotFoundError:
+        response_text = "Error: 'claude' command not found. Install Claude Code CLI first."
+    except Exception as exc:
+        response_text = f"Error: {exc}"
+
+    # Send response text as a single agent_message_chunk notification
+    _send_notification("session/update", {
+        "update": {
+            "sessionUpdate": "agent_message_chunk",
+            "content": {"text": response_text},
+        }
+    })
+
+    # Send completion response
+    _send_response(request_id, {"status": "completed"})
+
+
+_HANDLERS = {
+    "initialize": _handle_initialize,
+    "session/new": _handle_session_new,
+    "session/prompt": _handle_session_prompt,
+}
+
+
+def main():
+    # Log stderr so stdout stays clean for JSON-RPC
+    import logging
+    logging.basicConfig(
+        stream=sys.stderr,
+        level=logging.INFO,
+        format="%(asctime)s [%(levelname)s] claude-acp-bridge: %(message)s",
+    )
+
+    while True:
+        request = _read_request()
+        if request is None:
+            break
+
+        request_id = request.get("id")
+        method = request.get("method", "")
+        params = request.get("params", {})
+
+        handler = _HANDLERS.get(method)
+        if handler:
+            handler(request_id, params)
+        else:
+            _send_error(request_id, -32601, f"Method not found: {method}")
+
+
+if __name__ == "__main__":
+    main()