What problem does it solve? Writing software requirements in plain natural language produces ambiguity, untestable statements, and missed edge cases that surface only during implementation. This Skill provides a structured 4-phase methodology (Explore, Understand, Specify, Hand off) that turns vague feature ideas into formal, testable specifications before any code is written. ## Core Features & Use Cases - EARS Requirements: Writes requirements in Easy Approach to Requirements Syntax (ubiquitous, event-driven, state-driven, unwanted-behavior, optional patterns) so every statement is testable and unambiguous. - Diagram-Driven Thinking: Uses Mermaid mindmaps, sequence diagrams, state diagrams, ER diagrams, and C4 context diagrams to expose hidden assumptions and domain structure. - Scenario Table Generation: Derives happy, error, boundary, combinatorial, and adversarial (STRIDE-based) test scenarios directly from requirements and diagrams. - Ambiguity Detection: Applies a 9-type natural-language ambiguity checklist with red-flag word detection to catch vague requirements early. - Use Case: Before building a new file-upload feature, run the full spec process to produce EARS requirements, a state diagram of the upload lifecycle, a scenario table covering boundary and adversarial cases, and a beads epic ready for the planning phase. ## Quick Start Ask the AI to develop a formal specification for your feature using EARS notation and Mermaid diagrams, scaling formality to the risk of the task.
