Does this skill write source code?

No, Flow-Next Plan is strictly a planning tool. It generates epics and tasks in the .flow/ directory but does not modify your application's source code.

Why does it use a .flow/ directory instead of a TODO.md file?

The .flow/ directory provides a standardized, machine-readable format that prevents the AI from losing context or creating 'drift' that often occurs with free-form markdown files.

Can I use external LLMs for the review process?

Yes, the skill supports an 'export' mode for reviews, allowing you to validate your plans using external models like GPT-4 or Claude Web.

What is the primary purpose of Flow-Next Plan?

Its primary purpose is to create a reliable, structured blueprint for development before coding begins, ensuring tasks are properly sized and context-aware.

What is flowctl and do I need to install it?

flowctl is a utility used to manage task metadata. You do not need to install it globally; it is bundled within the plugin and called automatically by the skill.

Flow-Next Plan

Name: Flow-Next Plan
Author: gmickel

bygmickel

•

356

Productivity & Workflow

Generates structured implementation plans and granular task breakdowns using the Flow-Next architecture.

About

Flow-Next Plan is a specialized Claude Code skill designed to enforce a 'plan-first' development workflow. It transforms natural language feature requests or bug reports into high-fidelity epics and tasks stored within a dedicated .flow/ directory. By utilizing specialized research backends like RepoPrompt and grep, the skill scouts your repository to ensure plans respect existing patterns and reuse points. It acts as a product-minded architect, ensuring every task is sized for a single iteration and validated through optional multi-model review gates before a single line of code is written.

Key Features

Dual-mode repository scouting (Context-Scout via RepoPrompt or Repo-Scout via grep)
Strict .flow/ directory conventions to prevent task drift and state corruption
Bundled flowctl integration for machine-readable task state management
Automated epic and task tree generation from natural language or Flow IDs
Multi-model review gates supporting Codex, RepoPrompt, and external LLM exports
356 GitHub stars

Use Cases

Managing project state and task tracking directly within the terminal during AI coding sessions
Scaffolding repository-aware implementation plans that match existing code patterns
Breaking down complex feature requests into atomic, actionable development tasks

About

Key Features

Dual-mode repository scouting (Context-Scout via RepoPrompt or Repo-Scout via grep)
Strict .flow/ directory conventions to prevent task drift and state corruption
Bundled flowctl integration for machine-readable task state management
Automated epic and task tree generation from natural language or Flow IDs
Multi-model review gates supporting Codex, RepoPrompt, and external LLM exports
356 GitHub stars

Use Cases

Managing project state and task tracking directly within the terminal during AI coding sessions
Scaffolding repository-aware implementation plans that match existing code patterns
Breaking down complex feature requests into atomic, actionable development tasks