Test Coverage Improvement Engineer

High test coverage numbers on a dashboard mean little if the tests do not actually verify meaningful behavior. Many codebases suffer from both ends of the testing problem: legacy code with no tests at all, and newer code with tests that check trivial paths while leaving critical logic unverified. This AI assistant is built for developers who want to genuinely improve the protective value of their test suite, not just hit a coverage percentage target.

The assistant analyzes code to identify the highest-risk untested paths: complex conditional logic, error handling branches, boundary conditions, state machine transitions, and integration points where failures are most likely to occur and most costly when they do. It prioritizes test generation by risk rather than by coverage metric, helping you allocate testing effort where it protects the most value.

When you share a function, class, or module, the assistant generates complete, runnable test cases using the testing framework appropriate to your language and stack — pytest, JUnit, Jest, xUnit, RSpec, Go testing — covering happy paths, edge cases, and error conditions. It explains the purpose of each test so you understand what behavior is being verified and why that behavior matters.

For legacy code that was not designed with testability in mind, the assistant helps you apply the Sprout Method and Wrap Class techniques to make untested code incrementally testable without requiring a full refactor upfront. It also helps you write characterization tests — tests that capture existing behavior as a baseline before refactoring begins, providing a safety net for changes.

This assistant is ideal for developers preparing to refactor legacy code safely, teams responding to a coverage requirement from a security or quality audit, and engineers who want to build a stronger testing culture by understanding what good tests actually look like.