What problem does it solve?
Vague feature ideas in Elixir/Phoenix projects often lead to poorly scoped plans and rework. This Skill turns an unclear idea into a structured, coverage-scored requirements document through guided interviewing and targeted research, so downstream planning starts with concrete context instead of guesswork.
Core Features & Use Cases
- Adaptive Interview: Asks one question at a time across six coverage dimensions (What, Why, Scope, Where, How, Edge), scanning the codebase between questions to ground each question in real code.
- Research Cycles: Spawns up to two background agents (codebase patterns analyst and web researcher) to compare approaches with thesis/antithesis trade-off analysis.
- Structured Artifacts: Writes interview.md and research files to .claude/plans/{slug}/ that the phx-plan workflow consumes directly, skipping redundant clarification.
- Use Case: You type a vague request like "add some kind of notification system" and the Skill interviews you about scope, edge cases, and constraints, researches Phoenix notification patterns, then hands a complete requirements document to the planning step.
Quick Start
Ask the AI to run the phx-brainstorm skill with your feature idea, for example to brainstorm improving authentication security in your Phoenix app.
