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 agentmods add skills/xiaolai/mac-it-guy-pro/open-internetnpx skills add xiaolai/mac-it-guy-pro --skill open-internetgit clone --depth 1 https://github.com/xiaolai/mac-it-guy-proWrote 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/xiaolai/mac-it-guy-pro/open-internet)<a href="https://agentmods.dev/skills/xiaolai/mac-it-guy-pro/open-internet"><img src="https://agentmods.dev/badge/skills/xiaolai/mac-it-guy-pro/open-internet.svg" alt="Measured on agentmods" 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 | $0.00064 | $0.03208 |
| Opus 5 | $0.00032 | $0.01604 |
| Sonnet 5 | $0.00013 | $0.00642 |
| Haiku 4.5 | $0.00006 | $0.00321 |
Grade A, and why
open-internet scanned grade A with 0 findings 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 5d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Open Internet — what to build and why
Facts verified July 2026. Every price, protocol status, and provider claim in this skill and its references carries a date. Anything older than three months must be re-verified before you act on it — this field moves, and stale advice here costs the user real money.
Read references/legal-and-limits.md before anything else. Its boundaries are binding: personal and household use only, no resale, no "airport" operation. That is not a style preference — it is the line where the user's legal exposure changes category.
The rest of the detail lives in four more references, each loaded at the step that needs it: vps-buying.md (what to buy, what it costs, how to get a new IP), server-setup.md (build the server, both architectures), client-setup.md (connect the Mac and verify), and troubleshooting.md (diagnose in cheapest-first order). This file holds only the decisions that determine which of those you follow.
The one thing to get right: this is an architecture choice, not a protocol choice
Beginners (and most tutorials) argue about protocols. The decision that actually determines how a setup fails is whether the client connects to a CDN or directly to your server. The two resist different attacks, and no protocol changes that.
| CDN-fronted (VLESS/VMess + WS or httpupgrade + TLS behind Cloudflare) | Direct (VLESS + XTLS-Vision + REALITY on 443) | |
|---|---|---|
| What the censor sees | Ordinary HTTPS to a Cloudflare anycast IP | A genuine-looking TLS handshake to a borrowed hostname |
| Deep inspection | Measurably weaker than it looks. Xue et al. (USENIX Security 2024) detect the nested TLS handshake protocol-agnostically from burst sizes and round trips — 0.697 true-positive rate against shadowsocks-over-WebSocket-over-TLS at 0.05% false-positive rate. No evidence the GFW deploys this, but it is measured, not theoretical | Best available: Vision exists precisely to defeat that signature by splicing inner TLS records, uTLS matches a real browser, and active probes reach the genuine borrowed site |
| Origin IP exposure | Never exposed. China clients only ever touch Cloudflare | Fully exposed. Your server IP is the endpoint |
| IP blocking | Not applicable — blocking Cloudflare's pool is not politically cheap | No defence whatsoever. REALITY resists inspection, not a decision to null-route your address |
| Name/SNI blocking | This is its failure mode — your hostname can be blocked while Cloudflare stays up. Fix: new hostname | Not applicable — no hostname of yours is on the wire |
| Speed into China | Mediocre. No usable mainland Cloudflare presence on free plans; traffic hairpins to an overseas PoP and congests 20:00–23:00 | Determined by the server's route. A CN2 GIA box is dramatically better at peak |
| Moving parts | VPS + domain + CDN account + DNS + certificate + path config | VPS only. No domain, no certificate, no third-party account |
What ships with it
5 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.
- 5d ago First seen · 79 lines · 64 tokens per session scan A 4cab3e9ad1ba
open-internet is a skill published in the GitHub repository xiaolai/mac-it-guy-pro (5 stars, last pushed 1mo ago), licensed MIT. It adds 64 tokens to every session and 3,208 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
app-it
Turn a local project or a hosted web app (including a published Claude Artifact URL) into a macOS Dock-launchable .app. Use when the user wants a clickable Dock app, local app package, icon, App It install, hosted-URL or Artifact wrapper, or repeatable desktop launcher. Defaults to native Swift WebKit, shipped…
app-it-static
Turn a finished or buildable web app into a macOS Dock-launchable .app that serves built output instead of a dev server. Use when the user asks for app-it-static, a finished site/app launcher, a lightweight Dock app, or a dist/build/out bundle clickable from the Dock. Builds once, serves a snapshot via…
app-it-windows
Create Windows beta desktop launchers for local web projects: WPF + WebView2 .exe, Start Menu shortcut, .ico, warm server, quit cleanup. Use when the user asks for app-it-windows, a Windows app, .exe launcher, Start Menu shortcut, Windows package, or desktop icon. CI-guarded but untested on Windows hardware; never…
music
Apple Music in your terminal. Play tracks, route to AirPlay speakers and AirPods, search 100 million songs, build playlists, play radio stations, discover new music, favorite tracks, seek within a song, browse your listening history, browse your library by artist, album, or track. This is the plugin's only entry…
now
Show what's currently playing on this Mac (any app — Spotify, Apple Music, browsers). Use when the user asks what song/track/music is playing, who the artist is, or wants playback status.
output
Show or switch the Mac's audio output device (built-in speakers, AirPods, external displays, DACs). Use when the user asks where sound is playing, which output device is active, or wants audio routed to a different device.