AgentSkillsCN

reengine-playwright-agent

基于 Playwright 实现高效的人机交互自动化测试，配备强大的选择器、全程追踪、重试机制以及安全的人工介入防护措施。

SKILL.md
--- frontmatter
name: reengine-playwright-agent
description: Implement Playwright-based human-browser automation with robust selectors, tracing, retries, and safe human-in-the-loop gates.

Playwright Human-Browser Agent

Safety

  • •Never bypass CAPTCHA/2FA.
  • •Prefer API integrations over UI automation when possible.

Output artifacts

  • •traces on failure
  • •screenshots on failure
  • •structured logs

Deliverables

  • •a Playwright harness with persistent auth strategy
  • •a self-healing layer (retry taxonomy + fallbacks)
  • •an abstraction layer usable by MCP tools