Manages complex software development through hierarchical, outcome-based task trees and just-in-time planning.
The Discovery Tree Workflow (DTW) is a sophisticated system for planning and tracking work using a hierarchical breakdown of outcomes rather than linear steps. By utilizing 'beads' (bd), it allows developers and LLM agents to practice just-in-time planning, where the work tree grows organically as new requirements are discovered. This skill is specifically designed to facilitate agent coordination, allowing an orchestrator to define verifiable outcomes and hand them off to implementing agents for isolated, session-based execution that results in mergeable pull requests.
Key Features
01Visual status reporting for epics, tasks, and dependencies
020 GitHub stars
03Just-in-time planning to handle emergent requirements and complexity
04Hierarchical outcome tracking using the 'beads' (bd) system
05Support for competing implementation approaches for complex problems
06Agent handoff mechanism for session-based implementation
Use Cases
01Managing technical investigations where the roadmap emerges during the work
02Coordinating work between an orchestrator agent and specialized implementing agents
03Planning and executing multi-step feature developments with clear acceptance criteria
