🎭 Framework Upgradation — Claude Skills

A collection of Claude AI skills designed to automate the upgrade and migration of test automation frameworks to Playwright. These skills are powered by Claude's agent skill system and act as reusable, instruction-driven agents that can be invoked directly inside your workspace.

⚠️ Important Notice: The automated migration code generated by these skills has an estimated success rate of 50–60% only. Generated code should always be reviewed, tested, and manually corrected by a developer before being used in production. These skills are a starting point to accelerate migration — not a fully automated, zero-touch solution.

---

📁 Repository Structure

.claude/
└── skills/
    ├── java-to-playwright-typescript/
    │   └── SKILL.md          ← Selenium Java → Playwright TypeScript migration
    └── analyze-and-generate/
        └── SKILL.md          ← Analyze any framework → Generate Playwright in any language

---

🛠️ Skills Overview

1. java-to-playwright-typescript — Selenium Java → Playwright TypeScript

Trigger: Ask Claude to migrate a Selenium Java POM framework to Playwright TypeScript.

What it does:

• Performs a full codebase discovery (page objects, test classes, utilities, base classes, config, data providers, listeners, feature files)
• Reads .github/copilot-instructions.md first and applies all project conventions
• Maps every Java/Selenium concept to its idiomatic Playwright TypeScript equivalent
• Generates a complete, runnable Playwright TypeScript project with:
  • playwright.config.ts
  • package.json + tsconfig.json
  • Page objects using Playwright locator APIs (getByRole, getByLabel, getByTestId, etc.)
  • Fixtures-based base test setup
  • Full spec files — every @Test method migrated, zero omissions
  • .env.example with all extracted config values
• Enforces the READ → MAP → WRITE → RUN → VERIFY cycle for every test class, one at a time — no batching
• Runs stub detection scan after generation — no // TODO, no pass, no empty bodies allowed
• Executes per-spec verification to confirm every file launches without errors
• Produces a test count parity report confirming 100% test case migration

Key guarantees:

Guarantee	Detail
0 test omissions	Every @Test method must appear in the output
0 skeleton stubs	Every test body contains real Playwright actions and assertions
0 explicit waits	Thread.sleep() / WebDriverWait eliminated — Playwright auto-waits
100% logic migration	All business logic, conditionals, data setup, and teardown migrated
Per-spec execution	Every spec file is run and confirmed to launch cleanly

---

2. analyze-and-generate — Analyze Any Framework → Generate Playwright

Trigger: Ask Claude to analyze an existing test codebase and generate Playwright code in your preferred language.

What it does:

• Reads .github/copilot-instructions.md before any code generation — all project rules are respected
• Detects the target language from user request, Copilot instructions, or existing codebase (TypeScript, JavaScript, Python, Java, C#)
• Analyzes the full source project — framework, architecture, patterns, utilities, data strategy, reporters
• Produces a detailed Codebase Analysis Report before writing any code
• Maps every source concept to the correct Playwright equivalent for the target language
• Generates a complete project scaffold (config, fixtures, page objects, utilities, test specs)
• Enforces the READ → MAP → WRITE → RUN → VERIFY 5-step cycle for every test class
• Applies stub detection and per-spec execution verification

Supported source frameworks:

Source Framework	Detection
Selenium + TestNG / JUnit	✅
Selenium + pytest	✅
Cypress (JS/TS)	✅
WebdriverIO	✅
Puppeteer	✅
Robot Framework	✅
Cucumber (any driver)	✅
Protractor	✅
NightwatchJS	✅

Supported output languages:

Language	Test Runner
TypeScript	Playwright Test
JavaScript	Playwright Test
Python	pytest-playwright
Java	JUnit 5 / TestNG
C#	NUnit / MSTest

---

🔄 The Mandatory Migration Process

Both skills enforce a 5-step cycle that must be completed for every test class — one at a time, no batching:

READ  →  MAP  →  WRITE  →  RUN  →  VERIFY

Step	What happens
READ	Fully read the source class, all referenced page objects, helpers, base classes, data providers, and config
MAP	Produce an explicit mapping table converting every source element to its Playwright equivalent
WRITE	Write the complete spec file — every test body populated with real actions and assertions
RUN	Execute the spec file in isolation and confirm it launches without import/syntax/fixture errors
VERIFY	Log a ✅ COMPLETE entry with counts before moving to the next class

---

🚀 How to Use

Prerequisites

• Claude Code installed in your editor
• Your source test automation project open in the workspace

Setup

1. Clone this repository or copy the .claude/ folder into your project workspace:

git clone https://github.com/ankitcoforge/framework-upgradation.git
cp -r framework-upgradation/.claude /your-project/

2. (Optional but recommended) Add a Copilot instructions file with your project conventions:

.github/copilot-instructions.md

Invoke the Skills

Open Claude Code in your workspace and use natural language:

For Selenium Java → Playwright TypeScript:

Migrate my Selenium Java framework to Playwright TypeScript

For any framework → any language:

Analyze my Cypress project and generate equivalent Playwright tests in Python

Convert my WebdriverIO framework to Playwright TypeScript

Analyze my test suite and generate Playwright Java tests

---

✅ Quality Enforcement Built Into the Skills

Rule	Enforced
Read Copilot instructions before any code generation	✅
Full codebase analysis before writing	✅
Per-class READ → MAP → WRITE → RUN → VERIFY cycle	✅
No batching of classes	✅
No skeleton stubs (pass, // TODO, empty body)	✅
All business logic migrated (no simplification)	✅
All assertions migrated (none dropped or weakened)	✅
All data-driven rows preserved	✅
No sleep() / Thread.sleep() in output	✅
Per-spec execution verification	✅
100% test count parity report	✅

---

� Author

Ankit Srivastava

• GitHub: @ankitcoforge
• Repository: framework-upgradation

---

📄 License

MIT — © 2026 Ankit Srivastava