what is qe-ai fr?

QE-AI is a collection of quality engineering skills meant to be shared across different AI coding assistants, including Claude Code, OpenAI Codex, Gemini CLI, Cursor, Windsurf, and GitHub Copilot. The core idea is that each skill has one original file, written for Claude Code's skill format, and every other tool gets a thin adapter file that simply points back to that original, so updating a skill once updates it everywhere it is used. The repository includes over a dozen skills covering different parts of software testing work. Examples include auditing an existing test suite for fake or missing tests, generating test cases from requirements or specs, writing prompts specifically tuned for testing tasks, planning a test strategy for a feature, writing and maintaining automated API or UI tests, generating varied test data safely without using production data, analyzing bug reports and logs to find root causes, capturing screenshots across an entire user interface to spot visual problems, and managing reusable seed data for local development and end to end tests. There is also a skill for packaging a repeatable testing process into a new shared skill, following the same original plus adapter pattern. Beyond the skills themselves, the repository contains a training curriculum meant to help testers learn to apply AI to their work, organized by skill level from foundation through practitioner to builder. To use a skill with Claude Code, you can either clone this repository directly, since Claude Code will recognize the skills automatically, or copy a specific skill folder into your own project or into your global Claude configuration folder. Each other supported tool reads its own instruction file, such as AGENTS.md for Codex or GEMINI.md for Gemini CLI, and the repository ships a working example of each. If you would rather not keep the separate skill folder in your project, you can paste the skill's contents directly into your tool's instruction file instead, though you then need to manually keep it updated if the original skill changes. Adding a brand new skill follows a documented three step process of creating the file, linking it from every adapter, and listing it in the repository's skill table.