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 Wholiver/metis --skill metis-browsergit clone --depth 1 https://github.com/Wholiver/metisWrote 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/wholiver/metis/metis-browser)<a href="https://agentmods.dev/skills/wholiver/metis/metis-browser"><img src="https://agentmods.dev/badge/skills/wholiver/metis/metis-browser/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/wholiver/metis/metis-browser"><img src="https://agentmods.dev/badge/skills/wholiver/metis/metis-browser.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 | $0.00082 | $0.01722 |
| Opus 5.5 | $0.00033 | $0.00689 |
| Sonnet 5.5 | $0.00016 | $0.00344 |
| Haiku 4.5 | $0.00008 | $0.00172 |
Grade A, and why
metis-browser 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 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.
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.
Metis Browser
Use this skill before any built-in browser automation. Load it with read on this file or /skill:metis-browser.
When to use
- User asks to open, show, preview, navigate, click, type, fill, scroll, or screenshot a page in Metis Desktop Inspector / 内置浏览器.
- Local file preview: SVG, HTML, or other browser-viewable workspace files.
- Local web/dev preview (
localhost,127.0.0.1) that needs real browser interaction. - Inspect interactive UI state (buttons, forms, links) after code changes.
Do not use this skill for:
- Fetching documentation or API text → prefer
websearch/webfetch. - Opening an external OS browser → never substitute with
bashopen,open -a Safari,open -a "Google Chrome",xdg-open, orqlmanagewhenbrowser_*tools exist. - Desktop Computer Use against the Inspector webview → use
browser_*tools instead. - Binding local preview or test servers to port 5173 → that port is reserved for Metis Desktop Vite. Use another port (e.g.
--port 4173) and neverbrowser_navigateto:5173.
Required workflow
- Confirm
browser_*tools are listed for this session. If they are missing, report that the Desktop browser host is unavailable; do not invent Playwright/Chrome automation and do not use system browsers as a fallback for "内置浏览器". - In Build mode,
browser_snapshot,browser_take_screenshot, andbrowser_evaluateare readable and may run before admission.browser_navigate, mutatingbrowser_tabs(new/select),browser_click,browser_mouse,browser_fill,browser_type,browser_press_key, andbrowser_scrollare mutating: whenperformance_admitis available, admit first. Do not alwaysbrowser_navigatefirst with no admit.browser_tabslistis readable. - After admission (or when the tools are already allowed), prefer this order:
browser_navigate(opens/activates Inspector browser tab). Wait for the tool result; it must report a realurl/title, not a leftoverabout:blank. Localhost/127.0.0.1 reloads ignoring cache when already open.- Visual SVG/HTML/artwork:
browser_take_screenshotafter navigate returns. Snapshot of a drawing is usually empty and is not visual proof. - Interactive UI:
browser_snapshot(refs for buttons/inputs and visible canvas/video, pluspointerLocked/ viewport), thenbrowser_click,browser_mouse,browser_fill,browser_type,browser_press_key, orbrowser_scrollusing currentrefvalues. Snapshot again after every action (refs are session-local and become stale).
- Prefer
browser_take_screenshotwhen judging layout, color, animation, overlap, or SVG/HTML appearance. Preferbrowser_snapshotonly when you need refs to click or fill. - After editing local app code or an SVG/HTML file, reload with
browser_navigateto the same path before claiming the UI is fixed, then screenshot again.
What ships with it
2 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 Changed · +24 lines 2f609c71dfe8
- 1mo ago Changed · +1 lines c8fd3f940552
- 1mo ago Changed · -1 lines 01f6a27eca40
- 1mo ago First seen · 55 lines · 82 tokens per session scan A 875411d19197
metis-browser is a skill published in the GitHub repository Wholiver/metis (435 stars, last pushed 2d ago), licensed MIT. It adds 82 tokens to every session and 1,722 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-09-17.
Other skills, from other repositories
agent-browser
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a…
frontend-design
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI).…
generated-ui
Show an interactive calculator, simulator or comparison directly in the conversation when the user wants to adjust inputs and see results, such as splitting a bill, comparing plans or exploring parameters. Prefer it to writing an HTML file unless the user asks for a file, page or project code.
frontend-design
Produce intentional, responsive, accessible UI work and perform visual QA instead of generic component assembly.
webapp-testing
Start/reuse a local app, wait for readiness, inspect rendered state/console/network, act from observed selectors, and verify with evidence.
orienting-to-an-inherited-seat
Use when you have just been primed into an EXISTING seat through a planned handover — a different agent retired and handed you the seat plus its earned context — and you need a world model of what just happened to you. Covers how a handover differs from compaction and from a fresh launch, the…