Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Auto-detect network issues and force proxy usage with proxychains4. Use this skill when encountering connection timeouts, DNS failures, or blocked network access. Default proxy is http://127.0.0.1:9910
.claude/skills/aiskillstore-proxychains/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 119% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 94% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 54% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 86% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 23% | 0% |
自动使用代理解决网络问题 - 当遇到网络连接失败、超时或访问受限时，自动使用 proxychains4 强制走代理。
触发条件（自动应用）:
Claude 应该：
proxychains4http://127.0.0.1:9910bash# ❌ 原命令失败 curl https://github.com/user/repo # ✅ 使用代理重试 proxychains4 curl https://github.com/user/repo
场景 1: Git 操作失败
bash# 原命令 git clone https://github.com/user/repo.git # 自动改为 proxychains4 git clone https://github.com/user/repo.git
场景 2: Python pip 安装失败
bash# 原命令 pip install requests # 自动改为 proxychains4 pip install requests
场景 3: npm/yarn 安装失败
bash# 原命令 npm install package-name # 自动改为 proxychains4 npm install package-name
场景 4: wget/curl 下载失败
bash# 原命令 wget https://example.com/file.tar.gz # 自动改为 proxychains4 wget https://example.com/file.tar.gz
场景 5: Docker 拉取镜像失败
bash# 原命令 docker pull image:tag # 自动改为 proxychains4 docker pull image:tag
场景 6: SSH 连接失败
bash# 原命令 ssh user@remote-host # 自动改为 proxychains4 ssh user@remote-host
本地代理地址： http://127.0.0.1:9910
配置文件位置：
~/.proxychains/proxychains.conf (推荐)/etc/proxychains.conf (系统级)创建用户级配置（自动使用 127.0.0.1:9910）：
bashmkdir -p ~/.proxychains cat > ~/.proxychains/proxychains.conf << 'EOF' # Proxychains configuration strict_chain proxy_dns remote_dns_subnet 224 tcp_read_time_out 15000 tcp_connect_time_out 8000 [ProxyList] http 127.0.0.1 9910 EOF
bash# 设置代理环境变量 export PROXYCHAINS_SOCKS5_HOST=127.0.0.1 export PROXYCHAINS_SOCKS5_PORT=9910 # 使用 proxychains4 curl https://github.com
Proxychains 通过 LD_PRELOAD 机制拦截程序的网络调用：
重要限制：
IF 命令返回网络相关错误（timeout, connection refused, DNS failure）
THEN 自动使用 proxychains4 重试该命令IF 访问以下域名/服务：
- github.com
- raw.githubusercontent.com
- pypi.org
- npmjs.org
- registry.npmjs.org
- docker.io
- gcr.io
THEN 直接使用 proxychains4（不等待失败）IF proxychains4 命令也失败
THEN 提示用户：
1. 检查代理服务是否运行（127.0.0.1:9910）
2. 检查 proxychains 配置文件
3. 尝试其他代理地址bash# 测试代理是否可用 curl -x http://127.0.0.1:9910 https://www.google.com # 检查端口是否监听 netstat -tunlp | grep 9910 # 或 ss -tunlp | grep 9910
bash# 测试配置是否正确 proxychains4 curl https://ipinfo.io/json # 应该显示代理服务器的 IP，而不是本机 IP
错误 1: "proxychains: command not found"
bash# 安装 proxychains4 sudo apt install proxychains4 # Debian/Ubuntu sudo yum install proxychains-ng # CentOS/RHEL
错误 2: "timeout"
bash# 检查代理地址配置是否正确 cat ~/.proxychains/proxychains.conf | grep -A 2 "\[ProxyList\]" # 修改超时时间（在配置文件中） tcp_connect_time_out 15000 tcp_read_time_out 30000
错误 3: "can't read configuration file"
bash# 创建配置文件 mkdir -p ~/.proxychains cp /etc/proxychains.conf ~/.proxychains/proxychains.conf # 然后编辑配置
conf# ~/.proxychains/proxychains.conf strict_chain # 按顺序使用所有代理 [ProxyList] http 127.0.0.1 9910 socks5 127.0.0.1 1080
confdynamic_chain # 自动跳过死代理 [ProxyList] http 127.0.0.1 9910 http 127.0.0.1 8080 socks5 127.0.0.1 1080
confrandom_chain chain_len = 2 # 随机选择 2 个代理 [ProxyList] http 127.0.0.1 9910 socks5 127.0.0.1 1080 socks5 127.0.0.1 1081
bash# 使用自定义 DNS 通过代理解析 export PROXY_DNS_SERVER=8.8.8.8 proxychains4 curl https://example.com
references/proxychains.conf (完整示例)references/troubleshooting.mdreferences/quick-reference.md记住这些原则：
proxychains4proxychains4默认代理: http://127.0.0.1:9910
这个技能让 Claude 在遇到网络问题时自动使用代理，无需用户手动干预！
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 8,803 | 4,658 | -47% | 1 | 1 | 0% | 1,645 | 2,390 | +45% | 0 | 0 | — |
case-02 | fail→pass | 6,152 | 2,292 | -63% | 1 | 1 | 0% | 1,029 | 2,258 | +119% | 0 | 0 | — |
case-03 | fail→pass | 19,353 | 4,072 | -79% | 1 | 1 | 0% | 1,327 | 2,574 | +94% | 0 | 0 | — |
case-04 | fail→pass | 9,603 | 3,657 | -62% | 1 | 1 | 0% | 1,628 | 2,502 | +54% | 0 | 0 | — |
case-05 | fail→pass | 7,413 | 2,827 | -62% | 1 | 1 | 0% | 1,199 | 2,236 | +86% | 0 | 0 | — |
case-06 | fail→pass | 21,888 | 3,263 | -85% | 1 | 1 | 0% | 1,856 | 2,274 | +23% | 0 | 0 | — |
case-07 | fail→pass | 11,229 | 4,886 | -56% | 1 | 1 | 0% | 2,033 | 2,911 | +43% | 0 | 0 | — |
case-08 | fail→pass | 8,095 | 2,871 | -65% | 1 | 1 | 0% | 1,426 | 2,277 | +60% | 0 | 0 | — |
case-09 | fail→pass | 10,736 | 2,954 | -72% | 1 | 1 | 0% | 1,885 | 2,381 | +26% | 0 | 0 | — |
case-10 | fail→pass | 8,389 | 3,254 | -61% | 1 | 1 | 0% | 1,428 | 2,277 | +59% | 0 | 0 | — |
case-11 | pass→pass | 6,650 | 3,396 | -49% | 1 | 1 | 0% | 1,180 | 2,523 | +114% | 0 | 0 | — |
case-12 | fail→pass | 8,787 | 5,947 | -32% | 1 | 1 | 0% | 1,520 | 2,929 | +93% | 0 | 0 | — |
case-13 | pass→pass | 10,142 | 5,458 | -46% | 1 | 1 | 0% | 1,692 | 3,026 | +79% | 0 | 0 | — |
case-14 | fail→fail | 11,169 | 5,283 | -53% | 1 | 1 | 0% | 2,042 | 2,809 | +38% | 0 | 0 | — |
case-15 | pass→pass | 3,071 | 3,716 | +21% | 1 | 1 | 0% | 330 | 2,449 | +642% | 0 | 0 | — |
case-16 | pass→pass | 4,732 | 2,615 | -45% | 1 | 1 | 0% | 667 | 2,319 | +248% | 0 | 0 | — |
case-17 | pass→pass | 5,646 | 2,717 | -52% | 1 | 1 | 0% | 959 | 2,399 | +150% | 0 | 0 | — |
case-18 | fail→pass | 8,136 | 1,892 | -77% | 1 | 1 | 0% | 1,295 | 2,179 | +68% | 0 | 0 | — |
case-19 | pass→pass | 5,724 | 3,208 | -44% | 1 | 1 | 0% | 1,017 | 2,417 | +138% | 0 | 0 | — |
case-20 | pass→pass | 4,448 | 2,871 | -35% | 1 | 1 | 0% | 899 | 2,371 | +164% | 0 | 0 | — |
case-21 | pass→pass | 16,857 | 15,629 | -7% | 1 | 1 | 0% | 2,959 | 4,573 | +55% | 0 | 0 | — |
case-22 | fail→pass | 14,763 | 13,925 | -6% | 1 | 1 | 0% | 2,444 | 4,182 | +71% | 0 | 0 | — |
case-23 | pass→pass | 9,472 | 9,241 | -2% | 1 | 1 | 0% | 1,762 | 3,766 | +114% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 22 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +52 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.