What problem does it solve? Writing unit tests that genuinely cover branches and edge cases is tedious, and fixing failing tests without weakening assertions or deleting cases requires discipline. This Skill provides a structured workflow for both generating new unit tests and repairing failing ones while preserving test integrity. ## Core Features & Use Cases - Dual-Mode Operation: Generation mode writes tests covering normal flows, boundaries (null, extremes, thresholds), and exception branches, then executes them until they pass; repair mode diagnoses failure attribution (test code vs. production code) and applies minimal fixes, handing verification back to the user by default. - Failure Attribution Discipline: Distinguishes test-code problems (wrong expectations, missing mocks, order-dependent cases) from production defects, escalating cross-module issues to a bug-fix workflow instead of masking them. - Integrity Guardrails: Never weakens assertions, deletes cases, or changes correct expected values just to make tests green; isolates tests from real networks, clocks, randomness, and execution order using mocks and test doubles. - Use Case: A developer pastes a failing Maven test output and asks to make the tests pass. The Skill classifies each failure, fixes only the responsible code, reports the changes and production impact, and lets the user run the verification. ## Quick Start Ask the AI to generate unit tests for a specified class covering branches and edge cases, or paste failing test output and ask it to make those unit tests pass.