regression-testing

The Regression Testing skill provides a comprehensive framework for maintaining software integrity during active development. It is designed for QA engineers, software developers, and DevOps teams who need to ensure that code changes, dependency updates, and environment modifications do not introduce regressions into existing functionality. By shifting from 'test everything' to a risk-based, change-aware approach, teams can achieve high defect detection rates while drastically reducing execution times and computational costs.

• Intelligent Test Selection: Utilizes git-diff analysis and dependency mapping to identify only the tests impacted by specific code changes, significantly reducing suite execution time.

• Risk-Based Prioritization: Automatically ranks tests based on historical failure rates, business risk, and code impact to ensure the most critical paths are verified first.

• Automated Pipeline Integration: Seamlessly integrates into CI/CD workflows (GitHub Actions, GitLab CI, etc.) to orchestrate smoke tests, selective regression, and full pre-release test suites.

• Production Bug Feedback Loop: Automatically converts reported production bugs into regression test cases, ensuring that historical defects never resurface.

• Fleet Orchestration: Coordinates specialized agents including qe-regression-risk-analyzer, qe-test-executor, and qe-coverage-analyzer to manage complex regression environments.

• Use this skill after code commits, before major releases, or following infrastructure changes to maintain system stability.

• Inputs typically include pull request IDs, code diffs, and project-specific coverage maps; outputs provide optimized execution plans, pass/fail results, and risk coverage reports.

• Monitor suite health metrics such as flakiness and execution duration to optimize test maintenance; avoid skipping tests without proper impact analysis.

• The skill functions effectively across various technology stacks, including TypeScript, Python, Go, and Rust, using frameworks like Jest, Playwright, Cypress, pytest, and JUnit.

• Remember that while agentic selection is powerful, it is essential to verify test counts to ensure that coverage has not been inadvertently dropped during the optimization phase.