sprint

Multi-agent feature execution is managed by dividing large tasks into parallelizable waves, dynamically assigning agents to specific tasks, and automating communication. This approach streamlines complex refactoring by orchestrating the sprint with automated planning.

Wave-based execution in project planning divides a large feature or refactor task into smaller, manageable, and parallelizable waves. It allows teams to execute complex sprints incrementally rather than handling the entire feature simultaneously.

To start a sprint for a new feature, you use the sprint skill to initialize the process. It requires a well-defined project context, team structure, and artifact management to orchestrate both full sprints with research phases and lightweight wave executions.

Automated evidence capture works directly with artifact management by recording evidence of completed work at the artifact level. It ensures all task outputs within the assigned waves are systematically captured and documented during the sprint.

Before coordinating team collaboration for a feature sprint, you need a well-defined project context and an established team structure. These prerequisites ensure the sprint skill can effectively orchestrate agents and manage artifacts across the execution waves.

A full sprint with a research phase is used for complex features requiring preliminary investigation, while lightweight wave execution suits straightforward refactor tasks. Both require project context and team structure to properly orchestrate multi-agent collaboration.