1.9Self-Assessment Quiz

1. What fundamental limitation of traditional browser automation do agent skills address?
2. Do agent skills add new tools to the agent? If not, what do they add?
3. Name the six layers of the skill-based browser automation stack.
4. Why does Microsoft recommend the Playwright CLI over their own MCP server for coding agents?
5. Who pioneered the CLI-skill pattern, and who mainstreamed it?

Answers:

1. The gap between intent (what the test verifies) and implementation (exact selectors and steps).
2. No. Skills add procedural knowledge -- markdown instructions teaching the agent to use existing tools for a domain.
3. Test Definitions, AI Agent (hosting the skills), Bash Tool, playwright-cli, Playwright engine, Browser.
4. It is measurably more token-efficient: Microsoft's benchmark showed 4x fewer tokens per task (27k vs ~114k). Snapshots go to disk and are read on demand instead of streamed on every action.
5. Jason Huggins (creator of Selenium) with Vibium's vibe-check skill in 2025; Microsoft in 2026 with @playwright/cli and its install --skills flag.