Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add rfdiosuao/AgentSkill --skill openclaw-network-optimizergit clone --depth 1 https://github.com/rfdiosuao/AgentSkillWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/rfdiosuao/agentskill/openclaw-network-optimizer)<a href="https://agentmods.dev/skills/rfdiosuao/agentskill/openclaw-network-optimizer"><img src="https://agentmods.dev/badge/skills/rfdiosuao/agentskill/openclaw-network-optimizer/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/rfdiosuao/agentskill/openclaw-network-optimizer"><img src="https://agentmods.dev/badge/skills/rfdiosuao/agentskill/openclaw-network-optimizer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00102 | $0.02000 |
| Opus 5.5 | $0.00041 | $0.00800 |
| Sonnet 5.5 | $0.00020 | $0.00400 |
| Haiku 4.5 | $0.00010 | $0.00200 |
Grade A, and why
OpenClaw 网络优化助手 scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
bins: ["powershell", "curl"] How it starts
The opening of the file, as written. The whole thing — 199 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OpenClaw 网络优化助手
执行本 Skill 时,先建立证据,再提出小步可回滚修改,最后复测同一组指标。默认只读诊断;任何 DNS、代理、路由、网卡、服务顺序、WinHTTP、系统代理或 TUN 相关修改,都必须先说明风险、回滚方式和预期效果。
安全边界
- 锁定目标为速度、延迟、DNS 稳定性、AI 工具连通性和开发工具连通性。
- 保护 Clash Verge、Mihomo、Clash Meta、V2Ray、Tailscale、WireGuard、OpenVPN、Surge、Shadowrocket、企业 VPN 和 TUN 设备。
- 禁止把代理、VPN、TUN、虚拟网卡当成垃圾进程清理。
- 禁止删除网络配置、VPN 配置、路由规则、防火墙规则、代理配置文件或网络位置。
- 遇到远程桌面、SSH、Tailscale、RustDesk、企业 VPN 等远程访问迹象时,先询问再做任何可能断网的修改。
- 对每个修改保存原值,并在报告中写出回滚命令。
- 复测结果不稳定时,不宣称优化成功;优先保留可证实改善的改动,回退噪声较大的改动。
路由模式
选择并在报告中写明一种模式:
execute_direct: 只读诊断、报告生成、已明确批准的可回滚命令。plan_first: DNS 切换、系统代理修改、WinHTTP 修改、网卡指标修改、服务顺序修改、缓存刷新、Clash CLI 代理配置。clarify_first: 操作系统未知、远程连接风险不明、Clash/VPN 所属不明、用户要求改动代理节点或 TUN。
工作流程
1. 识别环境
先确认操作系统、主网络接口、默认网关、DNS、代理、路由和活跃代理进程。
Windows 优先使用 scripts/windows-network-snapshot.ps1 做只读快照:
powershell -ExecutionPolicy Bypass -File .\scripts\windows-network-snapshot.ps1 -ProxyPort 7897
若没有脚本,则手动执行同类命令:
Get-NetAdapter
Get-NetIPConfiguration
Get-DnsClientServerAddress -AddressFamily IPv4
Get-NetRoute -DestinationPrefix '0.0.0.0/0'
netsh winhttp show proxy
Get-ItemProperty 'HKCU:\Software\Microsoft\Windows\CurrentVersion\Internet Settings'
Get-NetTCPConnection -LocalPort 7897 -State Listen
2. 建立基线
记录同一时段的基线指标:
- 主接口、默认网关、DNS 服务器、系统代理、WinHTTP 代理、Clash/Mihomo 监听端口。
- 网关 ping、公网 ping、GitHub/OpenAI/国内站点 HTTP 时延。
- DNS 首次查询和二次查询耗时。
- Git、npm、pnpm、curl、Python/pip 是否读取代理。
- 是否走系统代理、CLI 环境变量代理、WinHTTP 代理或 Clash TUN。
将直连和代理路径分开测。Clash 端口常见为 7890、7897、7899,不要假设端口;从系统代理和监听端口确认。
3. 判断瓶颈
只输出 3-4 个证据最强的瓶颈:
- 本地链路问题:网关 ping 丢包、Wi-Fi 信号差、信道拥挤、驱动异常。
- DNS 问题:首次查询慢、路由器 DNS 波动、公共 DNS 明显更稳。
- 代理路径问题:直连不可用,代理可通但节点延迟高、HTTP 状态异常或规则未命中。
- CLI 代理缺失:浏览器可用,但 Git、npm、pnpm、curl、pip、OpenAI SDK 失败。
- WinHTTP 缺失:系统服务、部分 Windows 程序不走用户代理。
- Bufferbloat:带宽充足但加载时延大,优先建议路由器 SQM/CAKE/FQ-CoDel。
区分事实和假设。不要从单次 speedtest 或单个超时判断全局网络状态。
4. 提出修改
优先选择低风险、可回滚、局部生效的修改:
- CLI 代理配置:适合 Git/npm/pnpm/curl/OpenAI SDK 失败但浏览器可用。
- DNS A/B 测试:适合 DNS 首次查询慢或路由器 DNS 波动。
- DNS 缓存刷新:适合解析结果陈旧、域名刚切换、首次连接异常。
- WinHTTP 代理:适合 Windows 服务或特定系统组件不走用户代理。
- 网卡接口指标:适合默认路由被虚拟网卡抢占或路由优先级异常。
- 代理节点/规则建议:只读检查后提出建议;不要直接改 Clash 节点、规则、TUN 或配置文件。
What ships with it
10 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 199 lines · 102 tokens per session scan A 16d04cecd53b
OpenClaw 网络优化助手 is a skill published in the GitHub repository rfdiosuao/AgentSkill (16 stars, last pushed 2d ago), licensed MIT. It adds 102 tokens to every session and 2,000 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
incident-response-skill
A skill that uses object-style tool references with purpose descriptions and required flags, plus strict validation.
investigate-cloudflare
Investigate deployed Worker failures, regressions, and missing operational signals using Cloudflare Issues, logs, traces, SQL, or Observability MCP. Produce bounded evidence and verify recurrence against the deployed version.
tuturuuu-browser-vercel-debugging
Reproduce deployed Tuturuuu bugs and correlate browser evidence with Vercel logs.
python-project-deployment-troubleshooting
Read-only first troubleshooting workflow for Python project deployments, including systemd startup failures, dependency issues, encoding problems, permissions, proxy setup, and readiness gates.
gke-workload-scaling-troubleshooting
Diagnoses GKE HorizontalPodAutoscaler (HPA) failures — metrics showing as , FailedGetResourceMetric / FailedGetScale / FailedComputeMetricsReplicas events, missing Pod resource requests, custom/external metrics-pipeline breakage (FailedGetExternalMetric / FailedGetCustomMetric, unavailable metrics adapter…
gke-ai-troubleshooting-jobset-interruption
Diagnoses GKE JobSet interruptions, restarts, and preemptions for AI/ML training workloads autonomously. Use when troubleshooting JobSet restart loops, spot VM preemptions, node readiness failures, host VM issues, or coordinator worker crashes. Don't use for general GKE cluster creation, basic workload deployment, or…