Claude Dev Workflow - Phase-Based AI Development

Multi-agent AI development teams for iOS and web projects are managed through phase-based project orchestration, enforcing mandatory standards, ADRs, and phase gates to maintain governance across deployment sessions.

A phase-based AI workflow is a governance framework that uses phase gates to control project progression. You need phase gates when orchestrating multi-agent teams to ensure planning, analysis, and validation standards are met before advancing.

To start a multi-agent project, draft a Phase 0 PRD in docs/PRD.md, plan Phase 1 with Zen MCP planner, and deploy a 5-agent team for iOS or macOS development to initiate the orchestration.

Yes, you can use Zen MCP tools including planner, consensus, thinkdeep, analyze, precommit, and codereview to guide planning, decision making, and validation within the multi-agent AI development governance framework.

Session continuity across phase gates works by preserving knowledge-base documentation and managing session handoffs, enabling seamless transitions between phases for multi-agent AI development teams.

Yes, Architecture Decision Records (ADRs) are mandatory standards enforced by this workflow for macOS and web project orchestration, ensuring decisions are documented and validated across multi-agent phases.