behavioral-modes

Adaptive AI modes for coding workflows provide predefined behaviors for planning, debugging, reviewing, and releasing. They apply task-specific rules and output formats to keep the AI focused on the current project activity.

AI workflow modes include Brainstorm, Implement, Debug, Review, Teach, Ship, and Orchestrate. Each mode applies specific interaction styles and decision patterns to guide planning, coding, validation, and release workflows.

Start in Brainstorm mode to map project options, then switch to Implement mode to write production-ready code. This progression adapts AI behavior to match the evolving demands of your development workflow.

Multi-agent collaboration patterns like Explore, Plan-Execute-Critic, and Mental Model Sync are supported for complex tasks. These patterns coordinate behaviors across distinct modes during intricate coding, debugging, and orchestration workflows.

Automatic mode detection works for review and release workflows by identifying the task type and applying matching behavior rules. You can also use explicit switch commands to manually override the detected mode when needed.