proffesor-for-testing/agentic-qe

Agentic QE Fleet is an open-source AI-powered QA/QE platform designed for use with Coding Agents (works best with Claude Code) featuring specialized agents and skills to support testing activities for a product at any stage of the SDLC. Free to use, fork, build, and contribute. Based on the Agentic QE Framework created by Dragan Spiridonov.

What it solves

Agentic Quality Engineering (AQE) Fleet automates the tedious and complex parts of software testing and quality assurance. It solves the problem of maintaining high test coverage, identifying risky untested code paths, and fixing flaky tests across diverse tech stacks and coding agent platforms.

How it works

AQE operates as a fleet of specialized AI agents coordinated by a central "Queen Coordinator." This coordinator decomposes high-level quality tasks into smaller sub-tasks and assigns them to domain-specific agents (e.g., test architects, security scanners, or chaos engineers).

Key technical mechanisms include:

  • Intelligent Routing: A system called TinyDancer routes tasks to different LLM tiers (Haiku, Sonnet, Opus) based on complexity to balance cost and performance.
  • Pattern Learning: The system indexes successful test patterns and defect indicators in a persistent database, allowing it to reuse effective strategies across different sessions and projects.
  • MCP Server: It uses the Model Context Protocol (MCP) to integrate seamlessly with 11 different coding agent platforms like Claude Code, Cursor, and GitHub Copilot.
  • Quality Gates: It employs anti-sycophancy scoring to reject hollow or tautological tests.

Who it’s for

Software engineers, QA engineers, and development teams who want to integrate AI-driven testing into their existing CI/CD pipelines and coding agent workflows across multiple platforms.

Highlights

  • Massive Agent Fleet: Coordinates 60 specialized QE agents across 13 domains, including security, chaos engineering, and enterprise integrations (SAP/SOAP).
  • ** uma 75+ QE Skills:** A library of verified and validated skills ranging from TDD (London/Chicago) to mutation testing.
  • Multi-Platform Support: Works with 11 major coding agents including Cursor, Cline, and Windsurf.
  • Cross-Project Learning: Learns codebase patterns over time to improve test generation accuracy.
  • Cost Optimization: Supports local LLMs via Ollama and intelligent routing to minimize API spend.