Browser Test Skill

Run end-to-end browser tests on pages affected by a PR or branch changes using the agent-browser CLI.

Modes

• Manual (default): the user controls the dev server. Follow the steps below as written, including the headed/headless question.
• Pipeline (mode:pipeline): invoked by LFG or another automated runner. The run is unattended — never block on a question. Read references/pipeline-orchestration.md from this skill's directory and follow it; it overrides the free-port scan (step 4), dev-server startup (step 5), and the headed/headless question (step 6). It still uses the preferred port that step 4 computes.

Use agent-browser Only

Use the agent-browser CLI for every browser action in this skill — opening pages, clicking, filling forms, snapshots, screenshots — and use it exclusively. Do not use any other browser-automation tool, including a browser MCP integration, a built-in browser-control tool, Playwright, or Puppeteer. If the host offers several ways to drive a browser, always choose agent-browser.

• Claude Code: do not use Chrome MCP tools (mcp__claude-in-chrome__*).
• Codex: do not substitute unrelated browsing tools.

Workflow

1. Verify agent-browser Is Installed