What problem does it solve? Turning a vague feature idea into a rigorous, implementation-agnostic specification is error-prone: requirements get mixed with tech choices, uncertainties go unrecorded, and quality gates are configured inconsistently. This Skill enforces a disciplined SPECIFY protocol that produces a testable spec.md every time. ## Core Features & Use Cases - Structured Spec Generation: Creates .swarm/spec.md with numbered functional requirements (FR-001...), measurable success criteria (SC-001...), Given/When/Then scenarios, and explicit [NEEDS CLARIFICATION] markers — with zero implementation details. - Clarification Funnel: Classifies uncertainties (self_resolved, critic_resolved, research_needed, user_decision), consults a critic sounding board, and records resolved items as explicit assumptions, with a research timeout fallback. - QA Gate & Parallelism Selection: Runs a mandatory user dialogue covering eleven QA gates (reviewer, test_engineer, SAST, council modes, drift check, mutation testing, etc.), parallel coder count, commit frequency, and auto_proceed, persisting choices to .swarm/context.md. - External Plan Import: Reverse-engineers a spec from a user-provided implementation plan and validates it against task format rules. - Use Case: A user says "specify a notification preferences feature" — the Skill checks the codebase reality, delegates exploration and SME research, writes the spec, archives any prior version, and walks the user through gate selection before planning begins. ## Quick Start Ask the agent to specify a new feature, for example: "Specify a user notification preferences feature with email and push channels."