メニュー
Headless browser automation CLI. Open, browse, interact with, and screenshot web pages. Use when: opening websites, browser-based login, web app testing, taking screenshots, UI verification.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | agent-browser |
| description | Headless browser automation CLI. Open, browse, interact with, and screenshot web pages. Use when: opening websites, browser-based login, web app testing, taking screenshots, UI verification. |
| tags | ["browser","web","automation"] |
A headless browser automation tool by Vercel Labs. Open web pages, interact with elements, extract information, and take screenshots.
If not already installed:
npm install -g agent-browser && agent-browser install
npm install -g agent-browser: Install the CLIagent-browser install: Download Chrome for Testing (first time only; add --with-deps on Linux)Verify installation:
agent-browser --help
1. open <url> → Open a page
2. snapshot -i → Get interactive element snapshot (refs: @e1, @e2, etc.)
3. click/fill/scroll → Interact using refs
4. snapshot -i → Re-check state after interaction
5. screenshot → Save screenshot if needed
Important: Always run snapshot -i before interacting to get element refs.
agent-browser open <url>
agent-browser back
agent-browser forward
agent-browser reload
agent-browser close
agent-browser snapshot # Full page
agent-browser snapshot -i # Interactive elements only (recommended)
agent-browser snapshot -c # Compact view
agent-browser snapshot -d 3 # Depth-limited
agent-browser click @e1
agent-browser dblclick @e1
agent-browser fill @e2 "text" # Clear and type
agent-browser type @e2 "text" # Append text
agent-browser hover @e1
agent-browser check @e1 # Checkbox on
agent-browser uncheck @e1 # Checkbox off
agent-browser select @e1 "value" # Dropdown select
agent-browser press Enter # Key press
agent-browser scroll down 500 # Scroll
agent-browser scrollintoview @e1 # Scroll element into view
agent-browser wait 1500 # Wait milliseconds
agent-browser wait @e1 # Wait for element
agent-browser wait --text "Success" # Wait for text
agent-browser wait --load networkidle # Wait for network idle
agent-browser get title # Page title
agent-browser get url # Current URL
agent-browser get text @e1 # Element text
agent-browser get value @e1 # Input value
agent-browser screenshot # Current viewport
agent-browser screenshot path.png # Save to path
agent-browser screenshot --full # Full page
agent-browser screenshot --annotate # With element annotations
Save screenshots to your attachments/ directory and include in responses:
agent-browser screenshot ~/.animaworks/animas/{your_name}/attachments/screenshot.png
Find and interact with elements by role or label when refs are unclear:
agent-browser find role button click --name "Submit"
agent-browser find label "Email" fill "user@example.com"
agent-browser find text "Sign In" click
agent-browser state save auth.json # Save login state
agent-browser state load auth.json # Restore saved state
agent-browser --session s1 open site.com # Named session
agent-browser session list # List sessions
agent-browser open <url> --headed # Show browser window (GUI environments)
agent-browser console # Show console logs
agent-browser errors # Show error logs
agent-browser snapshot -i --json # JSON output
--headed for GUI display)AGENT_BROWSER_DEFAULT_TIMEOUT env var)Listen-only scribe workflow for Zoom meetings. Silently understand and record meeting transcripts delivered via RTMS, escalate only urgent matters to your supervisor, and post the report only after human approval once the meeting ends. Use when: receiving an inbox message with source=zoom (carrying a "[Zoom会議実況 チャンク#N …]" or "[Zoom会議終了 …]" header).
Zoom会議の聞き取り担当ワークフロー。RTMS経由で届く会議トランスクリプトを黙って理解・記録し、緊急事項のみ上司へエスカレーション、会議終了後に人間の承認を得てからレポートを投稿する。 Use when: source=zoom のinboxメッセージ（「[Zoom会議実況 チャンク#N …]」または「[Zoom会議終了 …]」ヘッダ付き）を受信したとき。
Zoom 회의 청취 담당 워크플로. RTMS를 통해 도착하는 회의 트랜스크립트를 조용히 이해·기록하고, 긴급 사항만 상사에게 에스컬레이션하며, 회의 종료 후 사람의 승인을 받은 뒤 리포트를 게시한다. Use when: source=zoom 인 inbox 메시지(「[Zoom会議実況 チャンク#N …]」 또는 「[Zoom会議終了 …]」 헤더 포함)를 수신했을 때.
Markdownスキルを作成するメタスキル。SKILL.mdのfrontmatterと本文、Progressive Disclosureとcreate_skillの手順を扱う。 Use when: 新規スキル追加、read_memory_file 向けの記述ルール確認、referencesやtemplates付きスキル生成が必要なとき。
cron.mdを正しいフォーマットで読み書きするスキル。定時タスクの追加・更新・削除手順を提供する。 Use when: cron.mdの編集、cron式の追加、LLM型・コマンド型タスクの追加・削除、定時ジョブのメンテナンスが必要なとき。
Skill for hiring and creating new Digital Anima in the AnimaWorks organization. Create a character sheet (Markdown) based on interviews, then use the CLI command (animaworks anima create) to generate identity/injection/permissions in bulk. After creation, self-configure via bootstrap. "Create a new employee", "Hire someone", "New employee", "Hiring", "Create Anima", "Recruitment", "Add team member", "Expand the team", "Add staff", "Create subordinate", "hire", "recruit", "team member"