⚙️ QA Engineer Agent

Reads your test_plan.md and writes plain English test cases + Playwright automation scripts — one file per module — acting as a Senior QA Engineer.

Part of the @agents-forge AIQA pipeline:

analyst → qa-planner → qa-engineer → qa-reviewer → qa-reporter

Install

npm install @agents-forge/qa-engineer

# Also install playwright-cli for token-efficient script generation
npm install -g @playwright/cli@latest
playwright-cli install --skills
npx playwright install chromium

Quick Start

# Run the full pipeline
npx @agents-forge/analyst https://my-app.com requirements.md
npx @agents-forge/qa-planner requirements.md test_plan.md
npx @agents-forge/qa-engineer test_plan.md --url https://my-app.com

# Then run your tests
npx playwright test --grep @smoke

Authentication

Auto-detected in this order — no config needed:

| Method | Setup | |---|---| | Anthropic API key | ANTHROPIC_API_KEY=sk-ant-... in .env | | Claude subscription | npm install -g @anthropic-ai/claude-code → claude login | | GitHub Copilot | gh auth login |

Features

• 📖 Reads test_plan.md and identifies all modules automatically
• 📝 Writes test-cases/<module>.md — professional plain English test cases
• 🎭 Writes tests/<module>.spec.ts — Playwright scripts using playwright-cli
• ⚙️ Scaffolds playwright.config.ts with HTML/JSON/JUnit reporters
• 🏷️ Tags every test: @smoke, @regression, @negative, @edge, @p0/p1/p2
• 🔐 Multi-provider auth — API key, claude login, or GitHub Copilot
• ⚡ Uses playwright-cli for token-efficient, reliable script generation

CLI Usage

# Basic — prompts for base URL
npx @agents-forge/qa-engineer ./test_plan.md

# With base URL provided
npx @agents-forge/qa-engineer ./test_plan.md --url https://www.saucedemo.com

Programmatic Usage

import { runQAEngineer } from "@agents-forge/qa-engineer";

await runQAEngineer({
  testPlanFile: "./test_plan.md",          // required
  baseUrl: "https://www.saucedemo.com",    // optional — prompted if omitted
  cwd: process.cwd(),                      // optional
});

Output Structure

test-cases/
├── login.md              ← plain English test cases
├── menu.md
├── checkout.md
└── cart.md

tests/
├── login.spec.ts         ← Playwright automation
├── menu.spec.ts
├── checkout.spec.ts
└── cart.spec.ts

playwright.config.ts      ← multi-browser config

Each .md file structure:

# Login — Test Cases
## Smoke Tests (@smoke @p0)
| ID | Title | Preconditions | Steps | Expected Result |
## Regression Tests (@regression @p1)
## Negative Tests (@negative @p1)
## Edge Cases (@edge @p2)

Each .spec.ts structure:

test.describe("Login — Smoke", () => { ... })    // @smoke @p0
test.describe("Login — Regression", () => { ... }) // @regression @p1
test.describe("Login — Negative", () => { ... })   // @negative @p1
test.describe("Login — Edge Cases", () => { ... }) // @edge @p2

Running Tests

npx playwright test                      # all tests
npx playwright test --grep @smoke        # smoke only
npx playwright test --grep @regression   # regression suite
npx playwright test --grep @p0           # P0 priority only
npx playwright test tests/login.spec.ts  # single module
npx playwright show-report               # HTML report

Prerequisites

| Requirement | Version | |---|---| | Node.js | >= 20.0.0 | | @playwright/cli | latest (global) | | Playwright skills | playwright-cli install --skills |

Environment Variables

| Variable | Description | |---|---| | ANTHROPIC_API_KEY | Anthropic API key | | BASE_URL | Override the base URL at test runtime | | QA_ENGINEER_MODEL | Override the AI model used |

Troubleshooting

| Error | Fix | |---|---| | Test plan not found | Run @agents-forge/qa-planner first | | playwright-cli not found | npm install -g @playwright/cli@latest | | Authentication failed | Set ANTHROPIC_API_KEY or run claude login | | Browser not installed | npx playwright install chromium | | Tests fail on CI | Set BASE_URL env var in your CI pipeline |

License

MIT