feat: 扩充 shell 命令白名单(含 containerd/ctr/nerdctl)，解除 AI Agent 轮数上限

- 新增 ~50 个 Linux 基础命令(du/top/iostat/sar/curl/tcpdump 等)
- 新增 containerd 系列: ctr (containers/images/tasks/snapshots),
  nerdctl (ps/logs/inspect/stats 等)
- ip 命令补充 maddr/monitor/tunnel/tuntap 只读操作
- _run_agent_loop 改为 while True 循环，max_steps=0 表示不限轮数
- main.py 传入 max_steps=0，AI Agent 持续诊断直到问题解决
- 更新 Agent system prompt 第7条：明确无轮数上限

backend/agent_tools.py

@@ -117,14 +117,24 @@ SHELL_READ_ONLY_COMMANDS: dict[str, set[str] | None] = {
     },
     "journalctl": None,  # 只读，任意参数 (-u, -n, --no-pager, --since 等)
     "crictl": {"ps", "inspect", "logs", "info", "version", "images", "stats"},
+    "ctr": {"containers", "images", "namespaces", "tasks", "snapshots", "info", "version", "plugins", "content", "leases"},
+    "nerdctl": {"ps", "logs", "inspect", "info", "version", "images", "stats", "top", "events", "system", "namespace", "network", "volume"},
     "docker": {"ps", "logs", "inspect", "stats", "version", "info", "images", "top"},
-    "df": None, "free": None, "uptime": None, "uname": None, "hostname": None,
-    "ip": {"addr", "address", "route", "link", "neighbor", "neigh", "rule"},
+    "df": None, "du": None, "free": None, "uptime": None, "uname": None, "hostname": None,
+    "ip": {"addr", "address", "route", "link", "neighbor", "neigh", "rule", "maddr", "monitor", "tunnel", "tuntap"},
     "ss": None, "netstat": None, "lsblk": None, "ls": None, "cat": None,
     "head": None, "tail": None, "wc": None, "grep": None, "egrep": None, "fgrep": None,
-    "ps": None, "mount": None, "lsmod": None, "lscpu": None, "lsof": None,
+    "ps": None, "pstree": None, "mount": None, "lsmod": None, "lscpu": None, "lsof": None,
     "dmesg": None, "ping": None, "nslookup": None, "dig": None, "getent": None,
     "timedatectl": None, "date": None, "stat": None, "file": None, "blkid": None,
+    "top": None, "htop": None, "iostat": None, "vmstat": None, "mpstat": None,
+    "pidstat": None, "sar": None, "nproc": None, "env": None, "which": None,
+    "readlink": None, "realpath": None, "find": None, "sort": None, "uniq": None,
+    "awk": None, "sed": None, "cut": None, "tr": None, "tee": None,
+    "basename": None, "dirname": None, "xargs": None, "bc": None, "expr": None,
+    "curl": None, "wget": None, "traceroute": None, "tracepath": None, "mtr": None,
+    "nc": None, "nmap": None, "telnet": None, "tcpdump": None, "ethtool": None,
+    "lspci": None, "lsusb": None, "dmidecode": None, "sysctl": None, "udevadm": None,
 }
 
 # etcdctl / etcdutl 由 classify_shell_command 中的专用块处理 (两段式命令)

backend/ai_agent.py

@@ -44,7 +44,7 @@ AGENT_SYSTEM_PROMPT = """你是 Kubernetes 自主诊断 Agent。你的首要职
 4. apply、delete、patch、scale、rollout、exec 等修改或交互命令必须暂停并请求人工批准。
 5. systemctl restart/stop/start 等系统修改命令必须暂停并请求人工批准。
 6. 不要猜测命令输出；必须依据实际工具结果继续判断。
-7. 最多执行有限轮次，优先使用 namespace 和资源名缩小范围。
+7. 只要问题没有完全解决（verified=true 且 remaining_issues=0），就要持续执行命令诊断，没有轮数上限。
 
 每轮必须只返回一个 JSON 对象，不要使用 Markdown 代码块：
 - 继续调查：{"action":"command","command":"kubectl ...","reason":"为什么执行"}
@@ -165,7 +165,10 @@ async def _run_agent_loop(
     question: str,
 ):
     invalid_response_count = 0
-    for step in range(start_step, max_steps + 1):
+    step = start_step
+    while True:
+        if max_steps > 0 and step > max_steps:
+            break
         yield {"type": "thinking", "step": step, "message": f"AI 正在规划第 {step} 轮诊断"}
         try:
             content = await _complete_chat(messages)
@@ -293,6 +296,9 @@ async def _run_agent_loop(
             {"role": "user", "content": "命令实际执行结果：\n" + json.dumps(result, ensure_ascii=False)},
         ])
 
+        step += 1
+
+    if max_steps > 0:
         yield {
             "type": "error",
             "message": (
@@ -300,6 +306,7 @@ async def _run_agent_loop(
                 "系统已停止继续执行以避免无限循环，请检查是否重复执行相同命令或 AI 服务响应异常。"
             ),
         }
+    # max_steps=0 表示不限轮数，不输出错误 — Agent 会持续执行直到完成或遇到致命错误
 
 
 async def resume_diagnostic_agent(pending: dict, result: dict):

backend/main.py

@@ -327,6 +327,7 @@ async def agent_diagnose_stream(question: str = "", execution_mode: str = "ai"):
             async for event in run_diagnostic_agent(
                 _last_report,
                 question=question,
+                max_steps=0,
                 execution_mode=execution_mode,
             ):
                 history_store.append_event(run_id, event.get("type", "event"), {k: v for k, v in event.items() if k != "type"})