WebMAC: Multi-Agent Collaborative Framework for Automated Web System Scenario Testing
arXiv·medium signal
WebMAC uses multiple specialized LLM agents collaborating to convert test scenarios into executable test scripts for web systems, moving beyond both manual test-script writing and single-agent approaches that lack the ability to coordinate across complex multi-step scenarios. The framework addresses a practical gap: LLM-based test generation tools often fail on scenarios requiring cross-page or cross-component coordination. Relevant to QA teams evaluating LLM-assisted testing pipelines.