multi-agent-orchestration

To parallelize tasks across multiple Claude agents, provide a PHASEx_TASKS.md file detailing epics and dependencies. The orchestration helper analyzes this task breakdown to determine agent count, assign epics, and generate the required coordination artifacts.

Multi-agent orchestration is the process of analyzing a phase's task breakdown to coordinate parallel work across multiple Claude agents. It extracts dependency graphs, determines agent count, and generates implementation prompts and coordination documentation.

You set up parallel work coordination by feeding your task breakdown into the orchestration helper. It reads your epic dependencies, determines the optimal agent count, and generates prompts and coordination documents to manage the parallel execution.

Yes, you need a PHASEx_TASKS.md file that details multiple epics and their dependencies. This file serves as the input for dependency analysis and graph extraction, enabling the generation of implementation prompts and coordination documentation.

The best way to manage parallel Claude agent assignments is using human-names-based collaboration patterns. The orchestration helper applies these patterns to improve readability and logging while generating prompts and coordination documentation for phase execution.

This orchestration approach applies when you have a PHASEx_TASKS.md with multiple epics and dependencies. If your phase has a single epic with no dependencies, the multi-agent parallelization and coordination artifacts are not necessary.