Agentic Quality Engineering Fleet (AQE) is an open-source AI-powered QA/QE platform designed to integrate with coding agents, offering specialized agents and skills for comprehensive testing throughout the Software Development Life Cycle (SDLC). It aims to automate test generation, identify coverage gaps, detect and fix flaky tests, and learn codebase patterns, thereby streamlining quality engineering processes and reducing AI costs for users.

How It Works

AQE employs a sophisticated agent coordination system orchestrated by a central "Queen Coordinator." This system manages over 60 specialized QE agents distributed across 13 domains, including test generation, security auditing, and chaos engineering. Tasks are decomposed, agents are spawned in parallel, and results are synthesized. Agents communicate via shared memory and consensus protocols. The platform features a pattern learning system that stores and reuses successful test strategies and defect indicators across sessions and projects, improving over time through experience replay and background consolidation. Intelligent model routing, managed by "TinyDancer," optimizes AI costs by directing tasks to appropriate LLM tiers (Haiku, Sonnet, Opus) based on complexity.

Quick Start & Requirements

• Primary install: npm install -g agentic-qe
• Initialization: Navigate to your project directory and run aqe init --auto. This command auto-detects the tech stack and configures the Multi-Cloud Platform (MCP) for immediate use with supported coding agents.
• Prerequisites: Requires Node.js and npm. Integration with specific coding agent platforms (e.g., Claude Code, GitHub Copilot, Cursor) is necessary, each with its own setup requirements. TypeScript 5.0+ is utilized.
• Links:
  • Release Notes: docs/releases/README.md
  • Changelog: CHANGELOG.md
  • Issues: https://github.com/proffesor-for-testing/agentic-qe/issues
  • Discussions: https://github.com/proffesor-for-testing/agentic-qe/discussions

Highlighted Details

• Orchestrates a fleet of over 60 specialized QE agents across 13 distinct domains.
• Supports integration with 11 coding agent platforms, including Claude Code, GitHub Copilot, and Cursor.
• Implements intelligent LLM routing to balance cost and quality across simple, moderate, and critical tasks.
• Features "Loki-mode" quality gates, such as anti-sycophancy scoring and automated injection of historical edge cases.
• Learns and leverages codebase patterns across projects for enhanced efficiency and consistency.

Maintenance & Community

The project is led by @proffesor-for-testing, with core development contributions from @fndlalit and @shaal, and architecture input from @mondweep. Community support and issue tracking are managed via GitHub Issues and Discussions. Sponsorship options are available for ongoing development.

Licensing & Compatibility

The project is released under the MIT License, which permits commercial use and integration into closed-source projects. It is compatible with various testing frameworks including Jest, Cypress, Playwright, Vitest, Mocha, pytest, and JUnit.

Limitations & Caveats

The README does not explicitly detail known limitations or alpha status. The setup complexity may increase with the number of integrated coding agent platforms. The effectiveness of advanced features like "Loki-mode" quality gates relies on the quality of learned patterns and agent performance.