What problem does it solve?
Prevents test-writing AI from producing thin happy-path suites by forcing test perspectives (equivalence, boundaries, combinations, state transitions, error guessing, and checklists) to be defined first, so boundaries, forbidden paths, and realistic failure modes are covered.
Core Features & Use Cases
- Perspective-first test planning: Turns “write tests” requests into an explicit plan of what to test before any code is generated.
- Classical QA techniques: Uses equivalence partitioning, boundary value analysis, decision tables, state transition testing, error guessing, and checklists to systematically drive coverage.
- Review and reinforcement: Supports requests to assess whether existing tests cover the right perspectives and to propose missing ones.
Quick Start
Ask an AI agent to use the qa-test-design skill to “enumerate test perspectives and then generate tests for X,” for example for a “login + role-based access” feature where both invalid states and edge cases must be considered.