What problem does it solve? Test suites accumulate design flaws — conditional logic, sleeps, missing assertions, external dependencies — that make tests fragile, slow, and falsely reassuring. This Skill performs a formal audit of test code using the testsmells.org 19-smell academic catalog, producing a severity-ranked report with citable smell names, exact locations, and concrete fixes. ## Core Features & Use Cases - Polyglot smell detection: Analyzes tests in .NET (MSTest/xUnit/NUnit/TUnit), Python (pytest/unittest), JS/TS (Jest/Vitest/Mocha), Java, Go, Ruby, Rust, Swift, Kotlin, PowerShell (Pester), and C++ (GoogleTest/Catch2). - Research-backed taxonomy: Maps every finding to a named smell from the literature (Assertion Roulette, Mystery Guest, Sleepy Test, Eager Test, Sensitive Equality, Magic Number Test, and more) with severity levels. - Calibration against false positives: Recognizes idiomatic patterns such as Go table-driven tests, pytest parametrize, mock verifications, and integration-test markers so legitimate code is not flagged. - Use Case: Point the Skill at a flaky test project and receive a dashboard of high/medium/low severity findings, code snippets showing each smell, remediation examples, and a prioritized fix plan. ## Quick Start Audit the tests in my project's test directory using the full academic test smell catalog and give me a severity-ranked report with fixes.
