What problem does it solve? Turning a rough project idea into clear, testable requirements is slow and error-prone, and AI-generated requirements often mix in implementation details or miss scope boundaries. This Skill produces structured requirements documents in EARS format that stay aligned with your project's steering context and record decisions and learnings for future sessions. ## Core Features & Use Cases - EARS Requirements Generation: Creates requirements with acceptance criteria written in EARS syntax, grouped into logical requirement areas with explicit inclusion and exclusion boundaries. - Lifecycle-Aware Routing: Reads spec.json workflow and artifact toggles to skip disabled phases, handle bugfix analysis mode, and respect design-first workflows before writing anything. - Review Gate and Feedback Capture: Runs a bounded requirements review gate before writing the file, and synchronously records directional user corrections as learnings and architectural choices as decisions. - Use Case: After running discovery for a new notifications feature, invoke this Skill to produce a requirements.md with numbered requirements, EARS acceptance criteria, and updated spec metadata, then get routed to the design phase. ## Quick Start Ask the assistant to generate requirements for your feature by running the kiro-spec-requirements skill with your feature name after discovery has created the spec directory.
