What problem does it solve? Projects often lack a clear strategic direction, causing misaligned decisions by contributors and AI agents. This Skill produces a VISION.md document that defines the project's purpose, success criteria, guiding principles, and boundaries so every change can be evaluated against a shared north star. ## Core Features & Use Cases - Structured Discovery: Analyzes the codebase, git history, and existing documentation to understand what the project actually is before writing anything. - Collaborative Clarification: Asks targeted questions based on project type (library, service, product, tool) to capture the user's intent and scope the document depth. - Actionable Vision Document: Generates a VISION.md with evaluable success criteria, principles with "which means" clauses, explicit boundaries, and quality standards. - CLAUDE.md Integration: Adds a reference to the vision in CLAUDE.md so all agents align their work with it. - Use Case: When starting a new open-source library, run this Skill to interview you about goals, analyze the repo, and produce a docs/VISION.md that future contributors and AI agents use to evaluate every feature proposal. ## Quick Start Ask the AI to create a vision document for this project and define its success criteria and guiding principles.