team-builder

To coordinate multiple AI agents for parallel execution, you can use an interactive menu to browse available agent personas, compose a team, and spawn the selected agents concurrently. The system aggregates their outputs into a unified report.

Organizing agent personas is best handled using a domain-subdirectory layout or a flat layout with prefix-based naming. This allows the system to automatically discover agent files, classify domains, and assemble teams by domain or prefix.

Aggregating outputs from parallel agents into a single report is handled automatically by the workflow orchestrator. After spawning the selected agents concurrently, it synthesizes their individual outputs into a unified report containing conclusions and next steps.

Yes, parallel agent execution works with markdown persona files in a flat directory layout. The system supports both flat and domain-subdirectory collections, allowing you to compose and spawn agent teams from either organizational structure.

If an agent fails during parallel workflow orchestration, the system handles the failure gracefully. It continues the execution process for the remaining agents and synthesizes a unified report with conclusions and next steps from the successfully completed outputs.

No, you do not need to manually write scripts to spawn agents from different domains. The system discovers agent files, classifies domains automatically, and provides an interactive menu to browse and compose teams on demand from available markdown personas.