Library › Book 1 › Test Agents -- Planner, Generator, Healer
Test Agents -- Planner, Generator, Healer
7.1🔒What Shipped in v1.56With v1.56 -- maturing through 1.61, current as of July 2026 -- Playwright gained three first-party test agents. Not a bolted-on AI…
7.2🔒One Command, Any LoopScaffolding is a single command, with a flag that carries most of the meaning:
7.3🔒The Artifact LayoutAfter scaffolding and a first pass of agent work, a repository looks like this:
7.4🔒Plan, Review, Generate, Heal -- the Division of LaborThe workflow as a pipeline:
7.5🔒The Healer: Self-Healing Done HonestlyThe healer's loop deserves precise statement, because one clause separates good architecture from dangerous marketing:
7.6🔒Limits an Architect Should NameInterviewers trust people who volunteer the limits of their own tools. Name these four unprompted:
7.7🔒Key Takeaways- Since v1.56, Playwright ships three first-party test agents: planner (app -> Markdown plans in specs/), generator (plans -> verified…
7.8🔒Q&ASelf-Assessment Quiz1. Name the three Playwright test agents and each one's input and output. 2. What does the --loop flag on init-agents select, and what…
7.9🔒Exercises[Beginner] Exercise 7.1: Scaffold the agents in a sandbox repo with npx playwright init-agents --loop=<your runtime>. Read every generated…
7.10🔒Career Translation- Deployed Playwright's first-party test agents (planner/generator/healer, v1.56+) under an existing coding-agent runtime via init-agents…
7.11🔒Q&AInterview Depth CheckPrompt: Explain the division of labor between the planner, generator, and healer -- and where you would place human review, and why exactly…