Getting it into your agent
There is no command for this one: it runs only inside a plugin, and the catalogue could not identify which plugin ships it. The source is linked below.
Wrote 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/aduermael/herm/webbrowser)<a href="https://agentmods.dev/skills/aduermael/herm/webbrowser"><img src="https://agentmods.dev/badge/skills/aduermael/herm/webbrowser.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 | $0.00033 | $0.03460 |
| Opus 5 | $0.00016 | $0.01730 |
| Sonnet 5 | $0.00007 | $0.00692 |
| Haiku 4.5 | $0.00003 | $0.00346 |
Grade A, and why
webbrowser 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 6d 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 — 279 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Webbrowser
Use this skill when a task needs web search, page browsing, browser interaction, screenshots, or a handoff where the user needs to see or operate the browser.
Prefer webbrowser over raw HTTP when the page needs JavaScript, logged-in state, form interaction, or visual inspection.
Cookbook (do this first)
You are reading this file because the skill was loaded into context. The webbrowser API is a global (like fs). Do not require("webbrowser").
Happy path — stay in the background and read the page yourself:
-- 1) Open (mobile layout + lean resources by default). Capture the return value.
local opened = webbrowser.open("https://www.google.com/search?q=barber+shop+near+Menlo+Park+CA")
print(opened) -- JSON: browser id, layoutMode, page snapshot
local browser = opened.browser
-- 2) Read content with page() — this is how you "see" the site without show().
local snap = webbrowser.page(browser, { fields = {"title", "url", "text", "actions"} })
print(snap)
-- 3) Answer the user from snap.page.text / actions. Do NOT call show() for this.
Rules that fix the common failures
- Never call
webbrowser.show()for ordinary search/browse. The browser runs offscreen.show()is only for user handoff (login, CAPTCHA, payment, or the user asked to see the window). - Always capture
open/createresults.local opened = webbrowser.open(url)thenopened.browser. Discarding the return loses the id and the first page snapshot. - Use
webbrowser.page(browser)to read sites. Do not inventcurrent_page,get,inspect, orsearch. If unsure,print(webbrowser.help())once. - Mobile by default. New browsers use phone viewport + mobile Safari UA (
layoutMode="mobile"). Switch the same instance withwebbrowser.set_layout(browser, "desktop")only when a site needs a desktop UI; switch back with"mobile"if helpful. - Prefer working in the main agent for one search + one
page()read. Use a sub-agent only for multi-page research, and give it this same recipe (globalwebbrowser,open→page, noshow).
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.
- 6d ago First seen · 279 lines · 33 tokens per session scan A 4ac95e91e53c
webbrowser is a skill published in the GitHub repository aduermael/herm (232 stars, last pushed 24d ago), licensed MIT. It adds 33 tokens to every session and 3,460 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
computer-use
Use Orca's computer-use CLI for OS/window-level inspection and input in visible local app windows. Use when a task must read or operate a native app or an external browser window (for example, Chrome, Edge, or Safari) or an app webview. Do not use for Orca's embedded browser or page-only browser automation. Use…
computer-use
Full desktop control on macOS, Windows, Linux, and HarmonyOS — accessibility-first observation and actions with pixel fallback, screenshots, zoom, screen recording, and switching between registered computers as a default.
webapp-testing
Start/reuse a local app, wait for readiness, inspect rendered state/console/network, act from observed selectors, and verify with evidence.
browse
Use the browse CLI for Browserbase browser automation, Browserbase cloud APIs, Browserbase Functions, templates, web fetch/search, diagnostics, and Browse.sh skill discovery/installation. Use when the user asks to navigate pages, inspect browser state, run local or remote browser sessions, manage Browserbase…
verify
Drive an engine app headlessly in a pty, record a video of the whole verification, and open a summary page (video + timeline + checks) with pixel open.
h5i
Use when browsing the web or a local app on behalf of a user, or when work should run inside a disposable, confined development box instead of on the host — opening a browser session and reading a page as an outline with @ref handles, auditing what that session actually reached, reviewing a pull request or any…