flow

Manage context-driven software development workflows with AI agents and persistent memory.

15|4|Updated Jan 20, 2026

Search Tags:#development workflow #TDD #knowledge management #AI coding #Beads #context-driven

Ready to run

npx skills add https://github.com/cofin/flow --skill flow-cofin

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: flow
Source: https://github.com/cofin/flow/tree/main/templates/antigravity/skills/flow
Command: npx skills add https://github.com/cofin/flow --skill flow-cofin

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill streamlines software development by providing a structured, context-aware workflow that integrates with AI coding assistants, ensuring consistency and deep project awareness.

Core Features & Use Cases

Spec-First Development: Define requirements and plans before coding.
Persistent Memory: Utilizes Beads for cross-session memory, surviving context compaction.
TDD Workflow: Enforces a Red-Green-Refactor cycle with coverage requirements.
Knowledge Capture: Elevates patterns and learnings across development cycles.
Use Case: Kickstart a new feature by initializing a flow, defining its spec, planning tasks, implementing them with AI assistance, and ensuring all learnings are captured for future use.

Quick Start

Initialize a new project with the flow setup command.

Frequently Asked Questions about flow

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I maintain persistent context for AI coding assistants across development sessions?▼

Persistent context for AI coding assistants is maintained using Beads for state management, which survives context compaction and ensures deep project awareness across multiple development sessions.

What is the best way to enforce a TDD workflow with AI agents?▼

Enforcing a TDD workflow with AI agents is achieved through a structured development flow that mandates a Red-Green-Refactor cycle and sets specific test coverage requirements before implementation.

Does this context-driven development workflow work with Claude Code and Gemini CLI?▼

Yes, this context-driven development workflow integrates directly with Claude Code, Gemini CLI, and Google Antigravity, using Beads to manage persistent memory and state across these tools.

How do I start a new feature using spec-first development with AI?▼

Spec-first development with AI begins by initializing a project flow, defining requirements and plans upfront, and planning tasks before moving into AI-assisted implementation and testing.

Can I capture and reuse development knowledge across different feature cycles?▼

Capturing development knowledge across cycles is supported by built-in knowledge management that elevates patterns and learnings, ensuring past insights are accessible for future AI-assisted feature development.

Why does my AI coding agent lose project context after session compaction?▼

AI coding agents lose project context during compaction because standard context windows are volatile; this workflow solves that by leveraging Beads to provide persistent memory that survives compaction.