Agent skill · Browser
agent-browser Grade A
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 button", "take a screenshot", "scrape data from a page", "test this web app", "login to
When it triggers
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.
Triggering is decided by the description field in the skill’s frontmatter — the agent matches your request against it. If your request never mentions that vocabulary, the skill will not load.
What it does
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 button", "take a screenshot", "scrape data from a page", "test this web app", "login to
Where it comes from
agent-browser is published by Vercel in vercel-labs/agent-browser, and the source file is vercel-labs/agent-browser/blob/main/skills/agent-browser/SKILL.md. CSkillBox indexes the frontmatter as written; we do not rewrite the author’s description or claim to have tested the skill.
Security scan
Grade A 100/100 — static scan of this SKILL.md, 2026-11-17.
No risky patterns found in the text: no piped shell downloads, recursive deletes, eval, credential reads or permission bypasses.
A static text scan, not a runtime audit — open the file before you run it. How grades work →
Install
npx skills add vercel-labs/agent-browser --skill agent-browserThis copies the skill into your agent’s skills folder. To target one agent explicitly, add -a claude-code (or your agent) to the command. Full install guide →
Before you install
A skill is instructions plus, often, scripts the agent will run. Read the SKILL.md first: check what it executes, which domains it calls, and whether it asks for permissions you are not comfortable granting. A directory entry is not an endorsement.
Community notes
Found a broken install? Know a skill we missed?
One line is enough: the skill name, what happened, and the command you ran. We read every note and fix the index.