mission-orchestrated

Multi-agent orchestration automates feature development by cycling three specialized agents—Pathfinder, Builder, and Inspector—through each task, ensuring focused exploration, implementation, and verification before recording completion.

The three-agent cycle is a task management mechanism where a Pathfinder agent explores the task, a Builder agent implements it, and an Inspector agent verifies the results, creating a deterministic workflow from start to close.

You manage feature lifecycle states by loading the feature, marking it in-progress, iterating through its tasks by spawning dedicated agents, and recording task completion to transition and ultimately close the feature.

Yes, this orchestration uses Beads for context management and bd for task management to implement deterministic task lifecycle commands, agent role templates, and explicit progress states for reproducible feature development.

The best way to track progress is by using Beads-context and structured agent outputs to maintain auditable task states and results, ensuring reproducibility throughout the feature lifecycle.

You should avoid multi-agent orchestration for simple, single-step modifications that do not require the deterministic Pathfinder, Builder, and Inspector cycle, as the overhead of managing task states and agent roles would outweigh the benefits.