trellis-brainstorm

To brainstorm requirements before AI coding, you need a structured discovery process that captures intent, scope, and constraints. This skill scaffolds a task directory, seeds a PRD, and uses iterative one-question-at-a-time Q&A to converge on an MVP.

Creating an MVP scope to prevent scope creep requires explicit convergence steps after exploring the design space. This skill guides you through expansion and convergence phases, surfacing edge cases and recording decisions to finalize a robust, bounded plan before implementation.

Generating a PRD from a rough task description involves seeding a document with initial intent and iteratively refining it. This skill triggers from your task description, auto-contextualizes the request, and researches technical choices to document focused decisions.

You do not need a complete specification before starting requirements discovery. This skill triggers from a basic task description and guides you through a repeatable workflow, asking focused questions one at a time to extract constraints and define the MVP scope collaboratively.

You can use this requirements discovery workflow for complex technical research and edge case analysis. It employs a research-first approach for technical choices and includes dedicated divergence and convergence phases to evaluate risks, future evolution, and edge cases for a robust plan.

The requirements discovery process asks questions one at a time to reduce back-and-forth and maintain focus. This one-question-at-a-time Q&A workflow systematically explores the design space, records decisions iteratively, and ensures convergence on a well-defined MVP before any coding begins.