This is a full-featured network debugging toolkit that wraps tshark, dumpcap, and native OS network utilities into a unified interface for embedded systems work. You get packet capture with BPF and display filters, offline pcap analysis, ping and traceroute tests, port scanning with banner grabbing, and live traffic statistics. It reads config from both skill-level and workspace-level JSON files with a sensible priority chain (CLI args beat workspace config beats state file beats defaults). The prompts trigger on mentions of Wireshark, tshark, Modbus TCP, EtherNet/IP, or phrases like "capture some packets" or "scan the ports". All scripts output structured JSON, and it handles the usual gotchas like missing Npcap drivers or permission errors with specific guidance. Solid choice if you're debugging embedded network stacks and want consistent tooling without switching between a dozen command line utilities.
npx -y skills add zhinkgit/embeddedskills --skill net --agent claude-codeInstalls into .claude/skills of the current project.
嵌入式网络通信调试工具,统一封装接口发现、抓包、离线分析、连通性测试、端口扫描和流量统计能力。
<skill-dir>/scripts/<skill-dir>/config.json(仅工具路径)<workspace>/.embeddedskills/config.json(网络参数)<skill-dir>/references/common_protocols.jsontshark (随 Wireshark 安装,需加入 PATH)dumpcap (随 Wireshark 安装)capinfosipconfig、ping、tracert、netstat、arp、nslookupskill/config.json)仅保留工具路径相关的环境级配置:
{
"tshark_exe": "tshark",
"capinfos_exe": "capinfos"
}
.embeddedskills/config.json)工作区下的 .embeddedskills/config.json 存放工程级网络配置:
{
"net": {
"interface": "",
"target": "",
"capture_filter": "",
"display_filter": "",
"duration": 30,
"timeout_ms": 1000,
"scan_ports": "",
"capture_format": "pcapng",
"log_dir": ".embeddedskills/logs/net"
}
}
--interface, --target 等) - 最高优先级.embeddedskills/config.json 中的 net 部分).embeddedskills/state.json 中的历史记录)连接和采集参数按优先级解析,脚本通过 CLI 参数接收覆盖值。若配置缺少必要项或连接失败,询问用户并引导修改配置。
tshark 是否可用;若不可用,提示用户安装 Wireshark(含 tshark)并确认已加入 PATH;若需要抓包,还需提示安装 Npcap 驱动;依赖缺失时终止执行并输出 status: error 及安装指引iface(列出网络接口)python <skill-dir>/scripts/net_iface.py [--filter <关键词>] [--tshark] [--json]
--tshark: 同时显示 tshark 抓包接口索引映射--filter: 按关键词筛选接口python <skill-dir>/scripts/net_capture.py [--interface <接口>] [--duration <秒>] [--capture-filter <过滤器>] [--display-filter <过滤器>] [--output <文件路径>] [--format <pcapng|pcap>] [--decode-as <规则>] [--json]
--interface: 抓包接口(覆盖配置)--duration: 抓包时长(覆盖配置)--capture-filter: BPF 抓包过滤器(覆盖配置)--display-filter: Wireshark 显示过滤器(覆盖配置)--output: 保存抓包文件路径--json: 输出 JSON Lines 格式(基于 tshark -T ek)--decode-as: 自定义解码规则python <skill-dir>/scripts/net_analyze.py <pcap_file> [--mode <summary|protocols|conversations|endpoints|io|anomalies|all>] [--filter <显示过滤器>] [--top <数量>] [--decode-as <规则>] [--export-fields <字段列表>] [--output <CSV路径>] [--json]
--mode all 输出全部分析维度python <skill-dir>/scripts/net_ping.py [--target <目标>] [--tcp <端口>] [--count <次数>] [--traceroute] [--concurrent <线程数>] [--timeout <毫秒>] [--json]
--target: 目标地址(覆盖配置)--tcp: TCP 连通性测试(指定端口)--traceroute: 执行路由追踪--timeout: 超时毫秒数(覆盖配置)python <skill-dir>/scripts/net_scan.py [--target <目标>] [--ports <端口范围>] [--timeout <毫秒>] [--banner] [--concurrent <线程数>] [--json]
--target: 目标地址(覆盖配置)--ports: 端口范围,如 '80,443,8000-8100'(覆盖配置)--banner: 尝试获取服务 Bannerpython <skill-dir>/scripts/net_stats.py [--interface <接口>] [--duration <秒>] [--display-filter <过滤器>] [--interval <秒>] [--mode <overview|protocol|endpoint|port>] [--json]
--interface: 抓包接口(覆盖配置)--duration: 统计时长(覆盖配置)--display-filter: Wireshark 显示过滤器(覆盖配置)所有脚本输出统一的 JSON 结构:
{
"status": "ok",
"action": "<子命令名>",
"summary": "<简要描述>",
"details": { ... }
}
错误时:
{
"status": "error",
"action": "<子命令名>",
"error": {
"code": "<错误码>",
"message": "<错误描述>"
}
}
capture --json 输出 JSON Lines,进度信息写入 stderr。
.embeddedskills/config.json需要查询嵌入式常用端口和协议映射时,读取 references/common_protocols.json。
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills