agent-browser
agent-browser enables Claude Code to interact with web browsers programmatically, handling clicks, text input, form submissions, and page navigation. Built for agentic workflows that need reliable browser automation, it bridges the gap between AI reasoning and real-world web tasks. Perfect for developers moving from manual scripting to autonomous agent-driven browser control.
agent-browser enables Claude Code to interact with web browsers programmatically by automating clicks, typing, form submissions, and page navigation. It bridges the gap between AI reasoning and real-world web tasks, allowing agents to perform browser interactions reliably without manual scripting. The tool is designed specifically for agentic workflows that need autonomous control over browser behavior.
AI-generated summary based on this skill's SKILL.md
Install
shanraisshan/claude-code-best-practice/agent-browser · repository language: HTML
git clone https://github.com/shanraisshan/claude-code-best-practice
cp -r claude-code-best-practice/.claude/skills/agent-browser ~/.claude/skills/agent-browserFrequently asked questions
AI-generated answers based on this skill's SKILL.md and metadata
How does agent-browser automate browser tasks with AI?
agent-browser enables Claude Code to interact with web browsers programmatically by automating clicks, typing, form submissions, and page navigation. It bridges the gap between AI reasoning and real-world web tasks, allowing agents to perform browser interactions reliably without manual scripting. The tool is designed specifically for agentic workflows that need autonomous control over browser behavior.
Can agent-browser click buttons and fill forms programmatically?
Yes, agent-browser supports clicking buttons and filling forms programmatically as core capabilities. You can automate form submission and data entry workflows, enabling agents to interact with web applications as if a human were manually entering data and clicking elements. This makes agent-browser ideal for automating repetitive web tasks and user workflows.
What is agent-browser's license and how can I use it?
agent-browser is released under the MIT license, which permits free use, modification, and distribution for both commercial and personal projects. The permissive MIT license makes agent-browser accessible for developers integrating browser automation into their agentic systems without licensing restrictions.
How does agent-browser handle authentication and browser state?
agent-browser maintains browser state across sessions, enabling it to handle authentication workflows automatically. This capability allows agents to log in to websites and preserve session information, making it possible to automate multi-step workflows that require persistent authentication without re-entering credentials for each interaction.
Can agent-browser extract data and take screenshots from web pages?
Yes, agent-browser supports both data extraction and visual capture. You can programmatically extract content and data from web pages while also capturing visual evidence via screenshots for debugging and verification. This combination makes agent-browser valuable for web scraping, testing, and monitoring tasks where both content and visual state matter.
What workflows is agent-browser designed for?
agent-browser is built for agentic workflows requiring reliable browser automation, including testing web applications by automating user workflows, navigating and interacting with web apps, and automating complex multi-step tasks. It's perfect for developers transitioning from manual scripting to autonomous agent-driven browser control, enabling AI systems to handle real-world web interactions independently.
SKILL.md
rendered from the published skill — quoted content, verbatim
Browser Automation with agent-browser
Core Workflow
Every browser automation follows this pattern:
- Navigate:
agent-browser open <url> - Snapshot:
agent-browser snapshot -i(get element refs like@e1,@e2) - Interact: Use refs to click, fill, select
- Re-snapshot: After navigation or DOM changes, get fresh refs
```bash agent-browser open https://example.com/form agent-browser snapshot -i
Output: @e1 [input type="email"], @e2 [input type="password"], @e3 [button] "Submit"
agent-browser fill @e1 "user@example.com" agent-browser fill @e2 "password123" agent-browser click
(truncated - see the full file via the links below)
Read as markdown · JSON record · Browse the source repository
File tree — 1 file
.claude/skills/agent-browser/SKILL.md