Defines structured project visions and verifiable goals for autonomous multi-session agent loops.
Soda Loop Vision is a specialized workflow skill designed to bridge the gap between high-level project ideas and concrete implementation plans. It facilitates an interactive, iterative dialogue to uncover hidden requirements, identify technical constraints, and decompose complex projects into a series of 3-10 verifiable goals. By generating a standardized VISION.md file and setting up dedicated loop directories, it provides autonomous agents with a clear, machine-verifiable roadmap for long-running tasks. This skill is particularly effective for developers using agentic workflows who need to ensure project scope is well-defined and success criteria are measurable before execution begins.
Key Features
01Interactive requirements discovery with iterative clarifying questions
02Decomposes project descriptions into concrete, verifiable goals with pass/fail conditions
03Automated loop directory management and slugified branch naming
04Support for importing existing plan files as seeds for goal elicitation
05Integrated Codex review for validating goal verifiability and constraint clarity
060 GitHub stars
Use Cases
01Defining the strategic blueprint for a complex multi-session coding task
02Translating vague project requirements into actionable, verifiable outcomes
03Establishing technical boundaries and scope for autonomous agent loops
