What problem does it solve?
This Skill removes the guesswork of interacting with complex websites by providing a reliable, agent-friendly workflow for navigation, element targeting, and verification using Playwright.
Core Features & Use Cases
- CLI-first browser automation: Use
@playwright/cli for token-efficient flows, with a snapshot → reference → interact → verify loop.
- Selective visibility & verification: Take screenshots and snapshots only when needed, while using
eval for text extraction to avoid unnecessary context cost.
- Auth persistence and session control: Use
--persistent profiles for logged-in experiences, and use named sessions (-s=) for parallel subagents without state collisions.
- Asset and artifact management: Direct snapshots/screenshots to
.pw-tmp/ and clean up stale artifacts to keep projects tidy.
- Fallback to MCP when appropriate: Use
@playwright/mcp for short exploratory sessions that benefit from richer page understanding.
Quick Start
Use the playwright Skill to complete a login-and-dashboard task by running CLI commands to open the target page in headed mode, taking a snapshot to find stable element refs, performing the required clicks/fills, and then taking a verification snapshot or screenshot.