mirror of
https://github.com/hotyue/IP-Sentinel.git
synced 2026-05-12 01:19:41 +08:00
229 lines
10 KiB
Bash
Executable File
229 lines
10 KiB
Bash
Executable File
#!/bin/bash
|
||
# ==========================================================
|
||
# IP-Sentinel: 深海声呐 (IP 质量全维异步检测模块 v4.0.0)
|
||
# ==========================================================
|
||
|
||
source /opt/ip_sentinel/config.conf
|
||
|
||
# ==========================================
|
||
# 1. 动态网络锚定与协议自适应 (专为多 IP / NAT 架构打造)
|
||
# ==========================================
|
||
DYNAMIC_IP_PREF="${IP_PREF:-4}"
|
||
PROBE_ARGS=("-y" "-j" "-f") # 默认注入: 自动确认、JSON格式、明文无掩码IP
|
||
|
||
# 强壮正则：支持 V4, V6 以及带有 [] 护甲的 V6 (兼容多 IP 站群机)
|
||
if [[ -n "$BIND_IP" && "$BIND_IP" =~ ^[0-9a-fA-F:\[\]\.]+$ ]]; then
|
||
RAW_BIND_IP=$(echo "$BIND_IP" | tr -d '[]')
|
||
# 严格探测物理网卡/虚拟 IP 存活状态，防止 IP 漂移导致探针彻底报错
|
||
if ip addr show 2>/dev/null | grep -qw "$RAW_BIND_IP"; then
|
||
# 恢复使用官方原生参数 -i，不再进行徒劳的底层劫持
|
||
PROBE_ARGS+=("-i" "$RAW_BIND_IP")
|
||
|
||
# 智能识别 V4 / V6，强制覆盖系统默认的 IP_PREF
|
||
if [[ "$RAW_BIND_IP" == *":"* ]]; then
|
||
DYNAMIC_IP_PREF="6"
|
||
elif [[ "$RAW_BIND_IP" == *"."* ]]; then
|
||
DYNAMIC_IP_PREF="4"
|
||
fi
|
||
fi
|
||
fi
|
||
|
||
# 补齐协议版本参数 (-4 或 -6)
|
||
PROBE_ARGS+=("-${DYNAMIC_IP_PREF}")
|
||
|
||
# 2. 静默拉取原始数据 (智能双重容灾: 官方主干优先防 RCE，短链双栈兜底纯 V6)
|
||
PROBE_SCRIPT="/opt/ip_sentinel/core/ip_probe.sh"
|
||
|
||
# [修复] 采用 -s 检查文件是否大于0字节，防止遗留 0 字节废弃文件
|
||
if [ ! -s "$PROBE_SCRIPT" ]; then
|
||
# 🛡️ 首选防线: 从 GitHub 官方拉取，杜绝短链 RCE 劫持风险
|
||
curl -sL -m 10 "https://raw.githubusercontent.com/xykt/IPQuality/main/ip.sh" -o "$PROBE_SCRIPT" 2>/dev/null
|
||
|
||
# 🚑 容灾防线: 如果拉取失败 (如纯 V6 机器无法解析 GitHub)，回退到 Cloudflare 双栈短链兜底
|
||
if [ ! -s "$PROBE_SCRIPT" ]; then
|
||
curl -sL -m 10 "https://IP.Check.Place" -o "$PROBE_SCRIPT" 2>/dev/null
|
||
fi
|
||
chmod +x "$PROBE_SCRIPT" 2>/dev/null
|
||
fi
|
||
|
||
# 封装打靶与清洗逻辑为函数，便于后续容灾重试
|
||
execute_probe() {
|
||
RAW_OUTPUT=$(timeout 180 bash "$PROBE_SCRIPT" "$@" 2>/dev/null)
|
||
JSON_DATA="{${RAW_OUTPUT#*\{}"
|
||
ESC=$(printf '\033')
|
||
JSON_DATA=$(printf "%s" "$JSON_DATA" | sed -e "s/${ESC}\[[0-9;]*[a-zA-Z]//g" -e "s/${ESC}[0-9;]*[a-zA-Z]//g" -e "s/x1b\\[[0-9;]*[a-zA-Z]//g" -e "s/x1b[0-9;]*[a-zA-Z]//g")
|
||
IP_ADDR=$(echo "$JSON_DATA" | jq -r '.Head.IP // empty' 2>/dev/null)
|
||
}
|
||
|
||
# 🚀 首轮实弹打靶 (带有物理网卡 BIND_IP 枷锁)
|
||
execute_probe "${PROBE_ARGS[@]}"
|
||
|
||
# 🚑 [容灾防线] 剥离枷锁抢救机制
|
||
# 如果打靶失败 (无 IP 回波)，且身上带着 -i 枷锁，极大概率是复杂路由 (如 WARP) 导致的死锁！
|
||
if [ -z "$IP_ADDR" ] && [[ "${PROBE_ARGS[*]}" == *"-i"* ]]; then
|
||
# 卸下 -i 物理枷锁，交由系统内核自主寻找最优路由，进行第二次抢救性探测
|
||
FALLBACK_ARGS=("-y" "-j" "-f" "-${DYNAMIC_IP_PREF}")
|
||
execute_probe "${FALLBACK_ARGS[@]}"
|
||
fi
|
||
|
||
if [ -z "$IP_ADDR" ]; then
|
||
curl -s -X POST "${TG_API_URL}" \
|
||
-d "chat_id=${CHAT_ID}" \
|
||
-d "parse_mode=Markdown" \
|
||
-d "text=❌ *深海声呐探测失败*
|
||
📍 节点：\`${NODE_ALIAS}\`
|
||
🌐 锁定IP：\`${PUBLIC_IP}\`
|
||
⚠️ *未收到有效回波。检测源超时或数据解析受阻。*" >/dev/null
|
||
exit 1
|
||
fi
|
||
|
||
[ -z "$IP_ADDR" ] && IP_ADDR="$PUBLIC_IP"
|
||
ASN=$(echo "$JSON_DATA" | jq -r '.Info.ASN // "Unknown"' 2>/dev/null)
|
||
ORG=$(echo "$JSON_DATA" | jq -r '.Info.Organization // "Unknown"' 2>/dev/null)
|
||
CITY=$(echo "$JSON_DATA" | jq -r '.Info.City.Name // "Unknown"' 2>/dev/null)
|
||
COUNTRY=$(echo "$JSON_DATA" | jq -r '.Info.Region.Name // "Unknown"' 2>/dev/null)
|
||
IP_TYPE=$(echo "$JSON_DATA" | jq -r '.Info.Type // "未知属性"' 2>/dev/null)
|
||
USAGE_TYPE=$(echo "$JSON_DATA" | jq -r '.Type.Usage.IPinfo // "未知场景"' 2>/dev/null)
|
||
|
||
# 3. 深度欺诈与信用评估 (各大权威库联查)
|
||
SCAM_SCORE=$(echo "$JSON_DATA" | jq -r '.Score.SCAMALYTICS // "0"' 2>/dev/null)
|
||
ABUSE_SCORE=$(echo "$JSON_DATA" | jq -r '.Score.AbuseIPDB // "0"' 2>/dev/null)
|
||
IPQS_SCORE=$(echo "$JSON_DATA" | jq -r '.Score.IPQS // "0"' 2>/dev/null)
|
||
IP2L_SCORE=$(echo "$JSON_DATA" | jq -r '.Score.IP2LOCATION // "0"' 2>/dev/null)
|
||
FRAUD_RISK=$(echo "$JSON_DATA" | jq -r '.Score.ipapi // "0%"' 2>/dev/null)
|
||
|
||
# [修复] 清洗 API 阻断返回的 null 值，保障面板整洁
|
||
[ "$SCAM_SCORE" == "null" ] || [ -z "$SCAM_SCORE" ] && SCAM_SCORE="N/A"
|
||
[ "$ABUSE_SCORE" == "null" ] || [ -z "$ABUSE_SCORE" ] && ABUSE_SCORE="N/A"
|
||
[ "$IPQS_SCORE" == "null" ] || [ -z "$IPQS_SCORE" ] && IPQS_SCORE="N/A"
|
||
[ "$IP2L_SCORE" == "null" ] || [ -z "$IP2L_SCORE" ] && IP2L_SCORE="N/A"
|
||
[ "$FRAUD_RISK" == "null" ] || [ -z "$FRAUD_RISK" ] && FRAUD_RISK="N/A"
|
||
|
||
# 代理/VPN 特征探针 (只要有一家认为是代理，就亮黄灯)
|
||
IS_PROXY="🟢 干净"
|
||
if echo "$JSON_DATA" | jq -e '.Factor.Proxy | to_entries | any(.value == true)' >/dev/null 2>&1 || \
|
||
echo "$JSON_DATA" | jq -e '.Factor.VPN | to_entries | any(.value == true)' >/dev/null 2>&1; then
|
||
IS_PROXY="🟡 疑似代理/VPN"
|
||
fi
|
||
|
||
# 4. 提取流媒体与 AI 解锁指标 (带解锁类型)
|
||
parse_media() {
|
||
local status=$(echo "$JSON_DATA" | jq -r ".Media.$1.Status // \"未知\"" 2>/dev/null)
|
||
local reg=$(echo "$JSON_DATA" | jq -r ".Media.$1.Region // \"\"" 2>/dev/null)
|
||
local type=$(echo "$JSON_DATA" | jq -r ".Media.$1.Type // \"\"" 2>/dev/null)
|
||
|
||
if [[ "$status" == *"解锁"* ]]; then
|
||
echo "🟢 ${reg} (${type})"
|
||
elif [[ "$status" == *"仅"* ]] || [[ "$status" == *"机房"* ]] || [[ "$status" == *"待支持"* ]]; then
|
||
# 捕捉 Netflix "仅自制"、ChatGPT "仅网页"、TikTok "机房" 等半残状态
|
||
echo "🟡 ${status} ${reg}"
|
||
elif [[ "$status" == *"屏蔽"* ]] || [[ "$status" == *"失败"* ]] || [[ "$status" == *"中国"* ]] || [[ "$status" == *"禁"* ]]; then
|
||
# 捕捉 "屏蔽"、"失败"、"禁会员"、"中国"(送中)
|
||
echo "🔴 ${status}"
|
||
else
|
||
echo "⚪ ${status}"
|
||
fi
|
||
}
|
||
|
||
NF_STAT=$(parse_media "Netflix")
|
||
YT_STAT=$(parse_media "Youtube")
|
||
DP_STAT=$(parse_media "DisneyPlus")
|
||
TK_STAT=$(parse_media "TikTok")
|
||
GPT_STAT=$(parse_media "ChatGPT")
|
||
APV_STAT=$(parse_media "AmazonPrimeVideo")
|
||
|
||
# 提取原生 JSON 里的原始状态用于底层隐写回传
|
||
RAW_NF_STAT=$(echo "$JSON_DATA" | jq -r '.Media.Netflix.Status // "Unknown"' 2>/dev/null)
|
||
RAW_YT_REG=$(echo "$JSON_DATA" | jq -r '.Media.Youtube.Region // ""' 2>/dev/null)
|
||
RAW_YT_STAT=$(echo "$JSON_DATA" | jq -r '.Media.Youtube.Status // "Unknown"' 2>/dev/null)
|
||
|
||
# 5. 邮局连通性与黑名单
|
||
PORT25=$(echo "$JSON_DATA" | jq -r '.Mail.Port25 // "false"' 2>/dev/null)
|
||
[ "$PORT25" == "true" ] && P25_TEXT="✅ 畅通" || P25_TEXT="❌ 封堵"
|
||
DNS_BLACK=$(echo "$JSON_DATA" | jq -r '.Mail.DNSBlacklist.Blacklisted // "0"' 2>/dev/null)
|
||
DNS_MARK=$(echo "$JSON_DATA" | jq -r '.Mail.DNSBlacklist.Marked // "0"' 2>/dev/null)
|
||
|
||
# 6. “送中” 逻辑判定
|
||
WARNING_MSG=""
|
||
# [修复] 官方 JSON 已经去除了方括号，直接匹配 CN 或者状态包含中国
|
||
if [[ "$RAW_YT_REG" == "CN" ]] || [[ "$RAW_YT_STAT" == *"中国"* ]]; then
|
||
# [修复] 采用 Bash 扩展转义 ($'...')，彻底解决直接打印 \n 字符的问题
|
||
WARNING_MSG=$'\n🚨 **[高危] 该节点已被 Google 判定为中国大陆 (送中)！**\n'
|
||
fi
|
||
|
||
# 7. 组装情报级 Markdown 战报
|
||
# 提取本地运行态版本与生成时间戳
|
||
LOCAL_VER="${AGENT_VERSION:-未知}"
|
||
CURRENT_TIME=$(date "+%Y-%m-%d %H:%M:%S")
|
||
# [核心修复] 抛弃本地残缺配置，直接提取探针刚刚实测拿到的真实出口 IP 拼接链接！
|
||
LINK_IP=$(echo "$IP_ADDR" | tr -d '[]')
|
||
|
||
REPORT="🎯 *IP-Sentinel 深海声呐报告*
|
||
📍 节点：\`${NODE_ALIAS}\`
|
||
🌐 地址：\`${IP_ADDR}\`${WARNING_MSG}
|
||
|
||
*🏢 物理身份与网络属性*
|
||
\`AS${ASN}\` | \`${ORG}\`
|
||
**定位:** \`${COUNTRY} - ${CITY}\`
|
||
**属性:** \`${IP_TYPE}\` | \`${USAGE_TYPE}\`
|
||
**探针:** ${IS_PROXY}
|
||
|
||
*🛡️ 欺诈雷达 (0为最优)*
|
||
• **Scamalytics:** \`${SCAM_SCORE}/100\`
|
||
• **AbuseIPDB:** \`${ABUSE_SCORE}/100\`
|
||
• **IPQS:** \`${IPQS_SCORE}/100\`
|
||
• **IP2Location:** \`${IP2L_SCORE}/100\`
|
||
• **IPAPI 风险率:** \`${FRAUD_RISK}\`
|
||
|
||
*🎬 核心业务解锁*
|
||
• **YouTube:** ${YT_STAT}
|
||
• **Netflix:** ${NF_STAT}
|
||
• **Disney+:** ${DP_STAT}
|
||
• **PrimeVideo:** ${APV_STAT}
|
||
• **TikTok:** ${TK_STAT}
|
||
• **ChatGPT:** ${GPT_STAT}
|
||
|
||
*✉️ 邮局与污染度*
|
||
• **25 端口出站:** ${P25_TEXT}
|
||
• **DNS 污染库:** 严重 \`${DNS_BLACK}\` | 轻微 \`${DNS_MARK}\`
|
||
|
||
_👉 [🔍 详细信用图谱直达 (Scamalytics)](https://scamalytics.com/ip/${LINK_IP})_
|
||
|
||
⏱️ \`${CURRENT_TIME}\` | ⚙️ \`v${LOCAL_VER}\`"
|
||
|
||
# [修复] 剥离显示层的 N/A，确保传给 Master 趋势数据库的是纯数字 (无效则记为0)
|
||
SAFE_SCAM_SCORE=$(echo "$SCAM_SCORE" | tr -cd '0-9')
|
||
[ -z "$SAFE_SCAM_SCORE" ] && SAFE_SCAM_SCORE="0"
|
||
|
||
# [v4.0.2 扩容] 提取 Google(基于YouTube) 和 ChatGPT 的原生状态
|
||
RAW_GOOG_STAT="${RAW_YT_REG:-$RAW_YT_STAT}"
|
||
[ -z "$RAW_GOOG_STAT" ] && RAW_GOOG_STAT="未知"
|
||
RAW_GPT_STAT=$(echo "$JSON_DATA" | jq -r '.Media.ChatGPT.Status // "未知"' 2>/dev/null)
|
||
|
||
# [修复] 废除会导致中文 UTF-8 字节被劈裂（产生乱码 ）的 awk 暴力截断。
|
||
# 原始状态文本极短（如"解锁"、"屏蔽"、"US"），只需洗掉隐形换行符即可安全传输。
|
||
S_GOOG=$(echo "$RAW_GOOG_STAT" | tr -d '\n\r ')
|
||
S_NF=$(echo "$RAW_NF_STAT" | tr -d '\n\r ')
|
||
S_GPT=$(echo "$RAW_GPT_STAT" | tr -d '\n\r ')
|
||
CB_DATA="svq|${NODE_NAME}|${SAFE_SCAM_SCORE}|${S_GOOG}|${S_NF}|${S_GPT}"
|
||
|
||
# 8. 挂载内联键盘并直送指挥部
|
||
JSON_PAYLOAD=$(jq -n \
|
||
--arg cid "$CHAT_ID" \
|
||
--arg txt "$REPORT" \
|
||
--arg cb "$CB_DATA" \
|
||
--arg cb_manage "manage:${NODE_NAME}" \
|
||
'{
|
||
chat_id: $cid,
|
||
text: $txt,
|
||
parse_mode: "Markdown",
|
||
disable_web_page_preview: true,
|
||
reply_markup: {
|
||
inline_keyboard: [
|
||
[{text: "📥 将本次体检录入趋势库", callback_data: $cb}],
|
||
[{text: "⚙️ 调出该节点控制台", callback_data: $cb_manage}]
|
||
]
|
||
}
|
||
}')
|
||
|
||
curl -s -X POST "${TG_API_URL}" -H "Content-Type: application/json" -d "$JSON_PAYLOAD" >/dev/null |