fix: 修复爬虫网络层、验证队列卡死及 API 500 错误

- 修复 BaseHTTPPlugin 连接池、并发控制、异常日志、超时策略
- 修复/增强 8 个爬虫插件的稳定性和 fallback 机制
- 清理 validation_tasks 表 4 万+ pending 任务，避免队列卡死
- 修复 app/api/main.py 缺失全局 app 实例导致的 500 错误
- 提升前端 Axios 超时到 120 秒，避免请求断开
- 修复插件统计持久化和调度器生命周期问题

app/plugins/base.py

@@ -1,7 +1,7 @@
 """通用 HTTP 爬虫基类 - 为基于 HTTP 请求的插件提供封装"""
 import random
 import asyncio
-import aiohttp
+import httpx
 from typing import List
 from app.core.plugin_system import BaseCrawlerPlugin
 
@@ -28,25 +28,39 @@ class BaseHTTPPlugin(BaseCrawlerPlugin):
             "Connection": "keep-alive",
         }
 
-    async def fetch(self, url: str, timeout: float = 10.0, retries: int = 3) -> str:
+    async def fetch(self, url: str, timeout: float = 15.0, retries: int = 2) -> str:
         """异步抓取指定 URL 的 HTML 内容"""
+        from app.core.log import logger
         headers = self.get_headers()
-        async with aiohttp.ClientSession(headers=headers) as session:
-            for attempt in range(retries):
+        transport = httpx.AsyncHTTPTransport(retries=0)
+        for attempt in range(retries):
+            async with httpx.AsyncClient(headers=headers, transport=transport, follow_redirects=True) as client:
                 try:
-                    async with session.get(
-                        url, timeout=aiohttp.ClientTimeout(total=timeout)
-                    ) as response:
-                        if response.status == 200:
-                            content = await response.read()
-                            encoding = response.get_encoding()
-                            if encoding == "utf-8" or not encoding:
-                                try:
-                                    return content.decode("utf-8")
-                                except UnicodeDecodeError:
-                                    return content.decode("gbk", errors="ignore")
-                            return content.decode(encoding, errors="ignore")
-                except Exception:
-                    pass
-                await asyncio.sleep(random.uniform(1, 3))
+                    response = await client.get(url, timeout=timeout)
+                    if response.status_code == 200:
+                        content = response.content
+                        encoding = response.encoding
+                        if encoding == "utf-8" or not encoding:
+                            try:
+                                return content.decode("utf-8")
+                            except UnicodeDecodeError:
+                                return content.decode("gbk", errors="ignore")
+                        return content.decode(encoding, errors="ignore")
+                    else:
+                        logger.warning(f"Fetch {url} returned status {response.status_code}")
+                except Exception as e:
+                    logger.warning(f"Fetch {url} failed (attempt {attempt + 1}/{retries}): {e}")
+                if attempt < retries - 1:
+                    await asyncio.sleep(random.uniform(1, 3))
         return ""
+
+    async def fetch_all(self, urls: List[str], timeout: float = 15.0) -> List[str]:
+        """并发抓取多个 URL，限制单个插件内部并发为 3"""
+        semaphore = asyncio.Semaphore(3)
+
+        async def _fetch_limited(url: str):
+            async with semaphore:
+                return await self.fetch(url, timeout=timeout)
+
+        tasks = [_fetch_limited(url) for url in urls]
+        return await asyncio.gather(*tasks)

app/plugins/fate0.py

@@ -13,27 +13,34 @@ class Fate0Plugin(BaseHTTPPlugin):
 
     def __init__(self):
         super().__init__()
-        self.urls = ["https://raw.githubusercontent.com/fate0/proxylist/master/proxy.list"]
+        self.urls = [
+            "https://raw.githubusercontent.com/fate0/proxylist/master/proxy.list",
+            "https://cdn.jsdelivr.net/gh/fate0/proxylist@master/proxy.list",
+        ]
 
     async def crawl(self) -> List[ProxyRaw]:
         results = []
+        # 顺序 fetch，带 fallback
         for url in self.urls:
             html = await self.fetch(url, timeout=30)
-            if not html:
+            if html:
+                break
+        if not html:
+            logger.warning(f"{self.display_name} 所有源均不可用")
+            return results
+        for line in html.split("\n"):
+            line = line.strip()
+            if not line:
+                continue
+            try:
+                data = json.loads(line)
+                ip = data.get("host")
+                port = data.get("port")
+                protocol = data.get("type", "http")
+                if ip and port:
+                    results.append(ProxyRaw(ip, int(port), protocol))
+            except Exception:
                 continue
-            for line in html.split("\n"):
-                line = line.strip()
-                if not line:
-                    continue
-                try:
-                    data = json.loads(line)
-                    ip = data.get("host")
-                    port = data.get("port")
-                    protocol = data.get("type", "http")
-                    if ip and port:
-                        results.append(ProxyRaw(ip, int(port), protocol))
-                except Exception:
-                    continue
         if results:
             logger.info(f"{self.display_name} 解析完成，获取 {len(results)} 个潜在代理")
         return results

app/plugins/ip3366.py

@@ -12,24 +12,29 @@ class Ip3366Plugin(BaseHTTPPlugin):
     name = "ip3366"
     display_name = "IP3366"
     description = "从 IP3366 网站爬取免费代理"
-    default_config = {"max_pages": 5}
+    default_config = {"max_pages": 3}
 
     def __init__(self):
         super().__init__()
         self._update_urls()
 
     def _update_urls(self):
-        max_pages = self.config.get("max_pages", 5)
+        max_pages = self.config.get("max_pages", 3)
         self.urls = [
             f"http://www.ip3366.net/free/?stype=1&page={i}" for i in range(1, max_pages + 1)
         ] + [
             f"http://www.ip3366.net/free/?stype=2&page={i}" for i in range(1, max_pages + 1)
         ]
 
+    def get_headers(self) -> dict:
+        headers = super().get_headers()
+        headers["Referer"] = "http://www.ip3366.net/free/"
+        return headers
+
     async def crawl(self) -> List[ProxyRaw]:
         results = []
-        for url in self.urls:
-            html = await self.fetch(url, timeout=15)
+        htmls = await self.fetch_all(self.urls)
+        for html in htmls:
             if not html:
                 continue
             soup = BeautifulSoup(html, "lxml")

app/plugins/ip89.py

@@ -1,3 +1,5 @@
+import asyncio
+import random
 import re
 from typing import List
 from bs4 import BeautifulSoup
@@ -35,6 +37,8 @@ class Ip89Plugin(BaseHTTPPlugin):
                     if re.match(r"^\d+\.\d+\.\d+\.\d+$", ip) and port.isdigit():
                         results.append(ProxyRaw(ip, int(port), "http"))
 
+            await asyncio.sleep(random.uniform(1, 2))
+
         if results:
             logger.info(f"{self.display_name} 解析完成，获取 {len(results)} 个潜在代理")
         return results

app/plugins/kuaidaili.py

@@ -1,4 +1,6 @@
 import re
+import asyncio
+import random
 from typing import List
 from bs4 import BeautifulSoup
 from app.core.plugin_system import ProxyRaw
@@ -16,22 +18,39 @@ class KuaiDaiLiPlugin(BaseHTTPPlugin):
 
     def __init__(self):
         super().__init__()
+        # 减少页数，降低被反爬概率，确保至少能拿到数据
         self.urls = [
-            f"https://www.kuaidaili.com/free/inha/{i}/" for i in range(1, 11)
-        ] + [
-            f"https://www.kuaidaili.com/free/intr/{i}/" for i in range(1, 11)
+            "https://www.kuaidaili.com/free/inha/1/",
+            "https://www.kuaidaili.com/free/intr/1/",
         ]
 
+    def get_headers(self) -> dict:
+        headers = super().get_headers()
+        headers["Referer"] = "https://www.kuaidaili.com/free/inha/"
+        headers["Accept"] = "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8"
+        headers["Accept-Encoding"] = "gzip, deflate, br"
+        headers["Accept-Language"] = "zh-CN,zh;q=0.9,en;q=0.8"
+        headers["Sec-Fetch-Dest"] = "document"
+        headers["Sec-Fetch-Mode"] = "navigate"
+        headers["Sec-Fetch-Site"] = "same-origin"
+        headers["Upgrade-Insecure-Requests"] = "1"
+        return headers
+
     async def crawl(self) -> List[ProxyRaw]:
         results = []
+        # 先访问首页预热会话，获取 cookie，降低被反爬概率
+        await self.fetch("https://www.kuaidaili.com/", timeout=10)
+        await asyncio.sleep(random.uniform(2, 4))
+
+        # 顺序请求免费代理页面
         for url in self.urls:
-            html = await self.fetch(url, timeout=15)
+            html = await self.fetch(url, timeout=10)
             if not html:
                 continue
             soup = BeautifulSoup(html, "lxml")
             table = soup.find("table")
             if not table:
-                logger.warning(f"{self.display_name} 未能找到表格，可能是触发了反爬")
+                logger.warning(f"{self.display_name} 未能找到表格，可能是触发了反爬: {url}")
                 continue
 
             for row in table.find_all("tr"):
@@ -44,6 +63,7 @@ class KuaiDaiLiPlugin(BaseHTTPPlugin):
                         protocol = "http"
                     if re.match(r"^\d+\.\d+\.\d+\.\d+$", ip) and port.isdigit():
                         results.append(ProxyRaw(ip, int(port), protocol))
+            await asyncio.sleep(random.uniform(5, 8))
 
         if results:
             logger.info(f"{self.display_name} 解析完成，获取 {len(results)} 个潜在代理")

app/plugins/proxylist_download.py

@@ -8,48 +8,87 @@ class ProxyListDownloadPlugin(BaseHTTPPlugin):
     default_config = {"max_pages": 5}
     name = "proxylist_download"
     display_name = "ProxyListDownload"
-    description = "从 ProxyListDownload API 获取代理"
+    description = "从 GitHub 公开代理列表获取代理"
 
     def __init__(self):
         super().__init__()
-        self.urls = [
-            "https://www.proxy-list.download/api/v1/get?type=http",
-            "https://www.proxy-list.download/api/v1/get?type=https",
-            "https://www.proxy-list.download/api/v1/get?type=socks4",
-            "https://www.proxy-list.download/api/v1/get?type=socks5",
+        # 首选 GitHub raw + fallback 备用源（jsdelivr CDN 或 ProxyScrape API）
+        self.sources = [
+            {
+                "primary": "https://raw.githubusercontent.com/komutan234/Proxy-List-Free/main/proxies/http.txt",
+                "fallbacks": [
+                    "https://cdn.jsdelivr.net/gh/komutan234/Proxy-List-Free@main/proxies/http.txt",
+                    "https://api.proxyscrape.com/v2/?request=get&protocol=http&timeout=10000&country=all&ssl=all&anonymity=all",
+                ],
+                "protocol": "http",
+            },
+            {
+                "primary": "https://raw.githubusercontent.com/komutan234/Proxy-List-Free/main/proxies/socks4.txt",
+                "fallbacks": [
+                    "https://cdn.jsdelivr.net/gh/komutan234/Proxy-List-Free@main/proxies/socks4.txt",
+                    "https://api.proxyscrape.com/v2/?request=get&protocol=socks4&timeout=10000&country=all",
+                ],
+                "protocol": "socks4",
+            },
+            {
+                "primary": "https://raw.githubusercontent.com/komutan234/Proxy-List-Free/main/proxies/socks5.txt",
+                "fallbacks": [
+                    "https://cdn.jsdelivr.net/gh/komutan234/Proxy-List-Free@main/proxies/socks5.txt",
+                    "https://api.proxyscrape.com/v2/?request=get&protocol=socks5&timeout=10000&country=all",
+                ],
+                "protocol": "socks5",
+            },
         ]
 
+    def _detect_protocol(self, url: str) -> str:
+        """根据 URL 判断协议（注意不要用 https:// 来判断）"""
+        if "socks4" in url:
+            return "socks4"
+        elif "socks5" in url:
+            return "socks5"
+        elif "/http.txt" in url or "protocol=http" in url:
+            return "http"
+        return "http"
+
+    def _parse_lines(self, html: str, protocol: str) -> List[ProxyRaw]:
+        """解析代理文本，统一处理 \r\n、\n 两种换行以及可能存在的空行"""
+        results = []
+        # 统一替换为 \n 后再分割
+        text = html.replace("\r\n", "\n").replace("\r", "\n")
+        for line in text.split("\n"):
+            line = line.strip()
+            if not line or ":" not in line:
+                continue
+            parts = line.split(":")
+            if len(parts) >= 2:
+                ip = parts[0].strip()
+                port = parts[1].strip()
+                if ip and port.isdigit():
+                    results.append(ProxyRaw(ip, int(port), protocol))
+        return results
+
     async def crawl(self) -> List[ProxyRaw]:
         results = []
-        for url in self.urls:
-            html = await self.fetch(url, timeout=30)
-            if not html:
+        # 并发请求所有 primary URL
+        primary_urls = [s["primary"] for s in self.sources]
+        primary_htmls = await self.fetch_all(primary_urls, timeout=15)
+
+        for idx, html in enumerate(primary_htmls):
+            source = self.sources[idx]
+            protocol = source.get("protocol") or self._detect_protocol(source["primary"])
+
+            if html and html.strip():
+                results.extend(self._parse_lines(html, protocol))
                 continue
 
-            # 根据 URL 判断协议
-            if "type=socks4" in url:
-                protocol = "socks4"
-            elif "type=socks5" in url:
-                protocol = "socks5"
-            elif "type=https" in url:
-                protocol = "https"
-            else:
-                protocol = "http"
-
-            lines = html.split("\r\n")
-            if len(lines) <= 1:
-                lines = html.split("\n")
-
-            for line in lines:
-                line = line.strip()
-                if not line or ":" not in line:
-                    continue
-                parts = line.split(":")
-                if len(parts) >= 2:
-                    ip = parts[0].strip()
-                    port = parts[1].strip()
-                    if ip and port.isdigit():
-                        results.append(ProxyRaw(ip, int(port), protocol))
+            # primary 返回空或仅空白字符，依次尝试 fallback
+            logger.warning(f"{self.display_name} 主源返回空，尝试 fallback: {source['primary']}")
+            for fallback_url in source["fallbacks"]:
+                fallback_html = await self.fetch(fallback_url, timeout=15)
+                if fallback_html and fallback_html.strip():
+                    fb_protocol = source.get("protocol") or self._detect_protocol(fallback_url)
+                    results.extend(self._parse_lines(fallback_html, fb_protocol))
+                    break
 
         if results:
             logger.info(f"{self.display_name} 解析完成，获得 {len(results)} 个潜在代理")

app/plugins/proxyscrape.py

@@ -1,4 +1,5 @@
 """ProxyScrape 测试爬虫 - 用于验证架构，支持全协议类型"""
+import asyncio
 from typing import List
 from app.core.plugin_system import ProxyRaw
 from app.plugins.base import BaseHTTPPlugin
@@ -19,39 +20,89 @@ class ProxyScrapePlugin(BaseHTTPPlugin):
 
     def __init__(self):
         super().__init__()
-        # 使用多个公开 GitHub 代理列表作为源，稳定性较差
+        # GitHub raw 源作为首选
         self.urls = [
             ("http", "https://raw.githubusercontent.com/monosans/proxy-list/main/proxies/http.txt"),
             ("https", "https://raw.githubusercontent.com/monosans/proxy-list/main/proxies/https.txt"),
             ("socks4", "https://raw.githubusercontent.com/monosans/proxy-list/main/proxies/socks4.txt"),
             ("socks5", "https://raw.githubusercontent.com/monosans/proxy-list/main/proxies/socks5.txt"),
         ]
+        # ProxyScrape 官方 API 作为 fallback
+        self.api_urls = {
+            "http": "https://api.proxyscrape.com/v2/?request=get&protocol=http&timeout=10000&country=all&ssl=all&anonymity=all",
+            "https": "https://api.proxyscrape.com/v2/?request=get&protocol=https&timeout=10000&country=all&ssl=all&anonymity=all",
+            "socks4": "https://api.proxyscrape.com/v2/?request=get&protocol=socks4&timeout=10000&country=all&ssl=all&anonymity=all",
+            "socks5": "https://api.proxyscrape.com/v2/?request=get&protocol=socks5&timeout=10000&country=all&ssl=all&anonymity=all",
+        }
+
+    def _parse_proxies(self, text: str, protocol: str) -> List[ProxyRaw]:
+        """解析 ip:port 每行的文本内容"""
+        proxies = []
+        for line in text.splitlines():
+            line = line.strip()
+            if not line or ":" not in line:
+                continue
+            parts = line.split(":")
+            if len(parts) >= 2:
+                ip = parts[0].strip()
+                port_str = parts[1].strip()
+                if port_str.isdigit():
+                    proxies.append(ProxyRaw(ip, int(port_str), protocol))
+        return proxies
 
     async def crawl(self) -> List[ProxyRaw]:
         results: List[ProxyRaw] = []
-        for protocol, url in self.urls:
+        protocols = [protocol for protocol, _ in self.urls]
+        urls = [url for _, url in self.urls]
+
+        # 1. 并发请求所有 GitHub raw 源，整体限时 10s，先完成的保留结果
+        tasks = [asyncio.create_task(self.fetch(url, timeout=12)) for url in urls]
+        done, pending = await asyncio.wait(tasks, timeout=10)
+        for task in pending:
+            task.cancel()
+        htmls = []
+        done_protocols = set()
+        for i, task in enumerate(tasks):
             try:
-                html = await self.fetch(url, timeout=30)
-                if not html:
-                    logger.warning(f"ProxyScrape {protocol.upper()} 返回空内容")
-                    continue
+                if task in done:
+                    htmls.append(task.result())
+                    done_protocols.add(protocols[i])
+                else:
+                    htmls.append("")
+            except Exception:
+                htmls.append("")
+                done_protocols.add(protocols[i])
 
-                count = 0
-                for line in html.splitlines():
-                    line = line.strip()
-                    if not line or ":" not in line:
-                        continue
-                    parts = line.split(":")
-                    if len(parts) >= 2:
-                        ip = parts[0].strip()
-                        port_str = parts[1].strip()
-                        if port_str.isdigit():
-                            results.append(ProxyRaw(ip, int(port_str), protocol))
-                            count += 1
+        fallback_protocols = []
+        for protocol, html in zip(protocols, htmls):
+            proxies = self._parse_proxies(html or "", protocol) if html else []
+            if proxies:
+                logger.info(f"ProxyScrape {protocol.upper()} GitHub raw 获取 {len(proxies)} 个代理")
+                results.extend(proxies)
+            else:
+                if protocol in done_protocols:
+                    logger.warning(f"ProxyScrape {protocol.upper()} GitHub raw 返回空或无效，将尝试 API fallback")
+                else:
+                    logger.warning(f"ProxyScrape {protocol.upper()} GitHub raw 请求超时，将尝试 API fallback")
+                fallback_protocols.append(protocol)
 
-                logger.info(f"ProxyScrape {protocol.upper()} 获取 {count} 个代理")
-            except Exception as e:
-                logger.error(f"ProxyScrape {protocol.upper()} 爬取失败: {e}")
+        # 2. 对 GitHub raw 失败的协议，并发请求 ProxyScrape API fallback
+        if fallback_protocols:
+            fallback_urls = [self.api_urls[p] for p in fallback_protocols]
+            try:
+                api_htmls = await asyncio.wait_for(
+                    self.fetch_all(fallback_urls, timeout=10), timeout=10
+                )
+            except asyncio.TimeoutError:
+                logger.warning(f"ProxyScrape API fallback 批量请求超时，跳过 {len(fallback_protocols)} 个协议")
+                api_htmls = [""] * len(fallback_protocols)
+            for protocol, api_html in zip(fallback_protocols, api_htmls):
+                proxies = self._parse_proxies(api_html or "", protocol) if api_html else []
+                if proxies:
+                    logger.info(f"ProxyScrape {protocol.upper()} API 获取 {len(proxies)} 个代理")
+                    results.extend(proxies)
+                else:
+                    logger.warning(f"ProxyScrape {protocol.upper()} API 返回空或无效")
 
         if results:
             logger.info(f"ProxyScrape 总计获取 {len(results)} 个代理")

app/plugins/speedx.py

@@ -18,13 +18,18 @@ class SpeedXPlugin(BaseHTTPPlugin):
             "https://raw.githubusercontent.com/TheSpeedX/SOCKS-List/master/socks4.txt",
             "https://raw.githubusercontent.com/TheSpeedX/SOCKS-List/master/socks5.txt",
         ]
+        self.fallback_urls = [
+            "https://cdn.jsdelivr.net/gh/TheSpeedX/SOCKS-List@master/http.txt",
+            "https://cdn.jsdelivr.net/gh/TheSpeedX/SOCKS-List@master/socks4.txt",
+            "https://cdn.jsdelivr.net/gh/TheSpeedX/SOCKS-List@master/socks5.txt",
+        ]
 
-    async def crawl(self) -> List[ProxyRaw]:
+    def _parse_htmls(self, htmls: List[str], urls: List[str]) -> List[ProxyRaw]:
         results = []
-        for url in self.urls:
-            html = await self.fetch(url, timeout=30)
+        for idx, html in enumerate(htmls):
             if not html:
                 continue
+            url = urls[idx]
 
             # 根据 URL 判断协议
             protocol = "http"
@@ -33,7 +38,7 @@ class SpeedXPlugin(BaseHTTPPlugin):
             elif "socks4" in url:
                 protocol = "socks4"
 
-            for line in html.split("\n"):
+            for line in html.splitlines():
                 line = line.strip()
                 if not line or ":" not in line:
                     continue
@@ -46,6 +51,16 @@ class SpeedXPlugin(BaseHTTPPlugin):
                     if not port.isdigit() or not (1 <= int(port) <= 65535):
                         continue
                     results.append(ProxyRaw(ip, int(port), protocol))
+        return results
+
+    async def crawl(self) -> List[ProxyRaw]:
+        htmls = await self.fetch_all(self.urls, timeout=15)
+        results = self._parse_htmls(htmls, self.urls)
+
+        if not results:
+            logger.warning(f"{self.display_name} GitHub 源全部返回空，尝试 jsdelivr fallback")
+            htmls = await self.fetch_all(self.fallback_urls, timeout=15)
+            results = self._parse_htmls(htmls, self.fallback_urls)
 
         if results:
             logger.info(f"{self.display_name} 解析完成，获取 {len(results)} 个潜在代理")

app/plugins/yundaili.py

@@ -1,6 +1,5 @@
 import re
 from typing import List
-from bs4 import BeautifulSoup
 from app.core.plugin_system import ProxyRaw
 from app.plugins.base import BaseHTTPPlugin
 from app.core.log import logger
@@ -12,41 +11,71 @@ class YunDaiLiPlugin(BaseHTTPPlugin):
     default_config = {"max_pages": 5}
     name = "yundaili"
     display_name = "云代理"
-    description = "从云代理网站爬取免费代理"
+    description = "从 GitHub 公开代理列表获取免费代理"
 
     def __init__(self):
         super().__init__()
+        # 主数据源：GitHub raw
         self.urls = [
-            f"http://www.ip3366.net/free/?stype=1&page={i}" for i in range(1, 6)
-        ] + [
-            f"http://www.ip3366.net/free/?stype=2&page={i}" for i in range(1, 6)
+            ("http", "https://raw.githubusercontent.com/mmpx12/proxy-list/master/http.txt"),
+            ("socks4", "https://raw.githubusercontent.com/mmpx12/proxy-list/master/socks4.txt"),
+            ("socks5", "https://raw.githubusercontent.com/mmpx12/proxy-list/master/socks5.txt"),
+        ]
+        # Fallback：jsdelivr CDN 加速
+        self.fallback_urls = [
+            ("http", "https://cdn.jsdelivr.net/gh/mmpx12/proxy-list@master/http.txt"),
+            ("socks4", "https://cdn.jsdelivr.net/gh/mmpx12/proxy-list@master/socks4.txt"),
+            ("socks5", "https://cdn.jsdelivr.net/gh/mmpx12/proxy-list@master/socks5.txt"),
         ]
 
-    async def crawl(self) -> List[ProxyRaw]:
-        results = []
-        for url in self.urls:
-            html = await self.fetch(url, timeout=15)
+    def _parse_htmls(self, htmls: List[str], url_mapping: List[tuple]) -> List[ProxyRaw]:
+        results: List[ProxyRaw] = []
+        for (protocol, _), html in zip(url_mapping, htmls):
             if not html:
-                continue
-            soup = BeautifulSoup(html, "lxml")
-            list_table = soup.find("div", id="list")
-            if not list_table:
-                continue
-            table = list_table.find("table")
-            if not table:
+                logger.warning(f"{self.display_name} {protocol.upper()} 返回空内容，可能网络受限或源已失效")
                 continue
 
-            for row in table.find_all("tr"):
-                tds = row.find_all("td")
-                if len(tds) >= 5:
-                    ip = tds[0].get_text(strip=True)
-                    port = tds[1].get_text(strip=True)
-                    protocol = tds[4].get_text(strip=True).lower() if len(tds) > 4 else "http"
-                    if protocol not in VALID_PROTOCOLS:
-                        protocol = "http"
-                    if re.match(r"^\d+\.\d+\.\d+\.\d+$", ip) and port.isdigit():
-                        results.append(ProxyRaw(ip, int(port), protocol))
+            count = 0
+            for line in html.splitlines():
+                line = line.strip()
+                if not line or ":" not in line:
+                    continue
+                parts = line.split(":")
+                if len(parts) < 2:
+                    continue
+                ip = parts[0].strip()
+                port_str = parts[1].strip()
+                if not re.match(r"^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}$", ip):
+                    continue
+                if not port_str.isdigit() or not (1 <= int(port_str) <= 65535):
+                    continue
+                final_protocol = protocol if protocol in VALID_PROTOCOLS else "http"
+                results.append(ProxyRaw(ip, int(port_str), final_protocol))
+                count += 1
+
+            if count:
+                logger.info(f"{self.display_name} {protocol.upper()} 解析完成，获取 {count} 个潜在代理")
+        return results
+
+    async def crawl(self) -> List[ProxyRaw]:
+        results: List[ProxyRaw] = []
+
+        # 顺序请求主源，避免某个 URL 卡住拖慢整体
+        for protocol, url in self.urls:
+            html = await self.fetch(url, timeout=12)
+            if html:
+                results.extend(self._parse_htmls([html], [(protocol, url)]))
+
+        # 主源为空时尝试 fallback（也顺序请求）
+        if not results:
+            logger.warning(f"{self.display_name} GitHub 主源全部返回空，尝试 jsdelivr fallback")
+            for protocol, url in self.fallback_urls:
+                html = await self.fetch(url, timeout=12)
+                if html:
+                    results.extend(self._parse_htmls([html], [(protocol, url)]))
 
         if results:
-            logger.info(f"{self.display_name} 解析完成，获取 {len(results)} 个潜在代理")
+            logger.info(f"{self.display_name} 总计解析完成，获取 {len(results)} 个潜在代理")
+        else:
+            logger.warning(f"{self.display_name} 未获取到任何代理")
         return results