Skip to main content

[Meetup] AI skills for QA that get you hired: Group task

  • April 9, 2026
  • 3 replies
  • 177 views

PolinaKr
Forum|alt.badge.img+6

GROUP ACTIVITY: Testing Agents Orchestration

1. Create Playwright Testing Agent
2. Create Backend Testing Agent
3. Create BDD Testing Agent

Orchestrate all these agents to work together in a coordinated manner to achieve comprehensive testing coverage for the application.

 

GROUP ACTIVITY: Testing Agents Orchestration with SKILLS

1. Create Playwright Testing Agent with SKILLS
2. Create Backend Testing Agent with SKILLS
3. Create BDD Testing Agent with SKILLS

3 replies

Forum|alt.badge.img+1
  • Specialist
  • April 9, 2026

 


Forum|alt.badge.img+1
  • Ensign
  • April 10, 2026

🚀 Just wrapped up this exciting project on AI Driven Multi Agent test orchestration! We built a custom multi-agent framework using GitHub Copilot to completely automate the testing pipeline for our Employee Management Application.

By utilizing a Test Orchestrator to coordinate three specialist agents (Backend, BDD, and Playwright), we achieved comprehensive three-layer coverage—resulting in 315 parallel tests passing across Chromium, Firefox, and WebKit in just 1.5 minutes.

📊 Check out the attached PDF project report for a deep dive into our agent architecture, skill design, and the key lessons we learned along the way! 👇

And here is the repo - Github Repo 

Adding our group members ​@mamtasharma0104  ​@richa 


gauravkhurana


3-Agent Orchestration with SKILL files + Live Reports

Built a coordinated multi-agent testing framework where a Test Orchestrator directs three specialist agents, each loaded with a SKILL.md domain knowledge file for their testing layer.

 

🏗️ Architecture:

Architecture Diagram

📊 Coverage — ~86 tests across 3 layers:

Layer Framework Tests
E2E Browser Playwright 34
Backend API Jest + Supertest 27
BDD Scenarios Cucumber + Playwright ~25

🔁 Phase-based CI (fail-fast strategy):

  1. Backend API tests first (fastest — confirms APIs work before spinning up browsers)
  2. BDD scenarios (validates business behaviour)
  3. Playwright E2E (full browser, slowest — runs last)

🌐 Live reports published separately to GitHub Pages:

🔗 Repository: https://github.com/gauravkhuraana/ActivityAI

What makes this unique: The 3 SKILL.md files act as a persistent, reusable knowledge base — any agent or developer can pick up the codebase and immediately know every selector, endpoint schema, step definition, and known pitfall without reading source code.

 


Team Members

Chamali J. Perera | LinkedIn

Viswanathan Sekar | LinkedIn

Piyush Jeswani | LinkedIn

Gaurav Khurana | LinkedIn