What problem does it solve? AI coding sessions lose context between runs, frontend and backend drift apart, and nobody can say what is actually finished. Archflow structures AI-assisted development into explicit lifecycle phases with file-based handoffs, so project state survives sessions and agents work from shared contracts instead of chat history. ## Core Features & Use Cases - Phase-based lifecycle management: Guides projects through strategy, design, API architecture, implementation, quality, launch, and enhancement phases, each with its own instruction file and approval gate. - Release and backlog management: Tracks stories across roadmap, release, backlog, and history YAML files with a readiness pipeline from backlog stub to shipped release. - Specialized agent orchestration: Dispatches domain agents (product-strategist, ux-designer, api-contract-architect, ui-engineer, qa-engineer, and others) that communicate through repository artifacts rather than messages. - Design system and API contract enforcement: Requires every UI-producing agent to follow the project's chosen design system file and binds frontend and backend work to a single API contract. - Use Case: A developer with an existing codebase runs /archflow:onboard to audit the stack and import context, then uses /archflow:release and /archflow:feature to plan and build stories one at a time, with QA and acceptance review before anything merges. ## Quick Start Ask the agent to run /archflow:status to see where the project stands and what to do next, or /archflow:init to set up a new project.