behavioral-modes

You can adapt AI behavior for coding tasks by switching between six operational modes that adjust reasoning style, questioning, and output structure to match specific planning, debugging, or code review scenarios.

Multi-agent collaboration patterns like Plan-Execute-Critic and Mental Model Sync coordinate outputs across agents, enabling structured code reviews by separating execution, discovery, and critique across distinct AI workflows.

Automatic mode detection triggers from specific prompt phrases, shifting the assistant into planning or implementation modes to enforce constraints like concise production code without requiring manual configuration.

Yes, you can use explicit slash commands like /implement to manually switch modes, enforcing strict constraints for deployment, teaching, or coding tasks across single-agent and multi-agent workflows.

Verbose code often occurs when the AI lacks task context; using the IMPLEMENT mode enforces mode-specific constraints that generate concise production code and tests instead of unhelpful verbosity.

Divergent ideation is needed during early planning phases; the BRAINSTORM mode adjusts the AI's reasoning style to provide multi-option ideation for exploring software architecture and feature implementation.