What problem does it solve?
Coordinating multiple development stories in parallel often leads to file conflicts, broken dependency order, and manual merge chaos. This Skill plans a wave of stories as a dependency DAG, partitions them into non-conflicting batches, and dispatches each story through the full-sdc development cycle with a controlled fan-in and merge handoff.
Core Features & Use Cases
- Deterministic Wave Planning: Builds a story DAG via topological sort on
depends_on and partitions batches by File List overlap using the aiox wave CLI.
- Batch Dispatch with full-sdc: Runs each story through the full-sdc execute protocol, in parallel when multi-agent spawning is available or sequentially as a fallback.
- Fan-in and Merge Handoff: Re-checks file conflicts across branches before merge and hands off PR/merge exclusively to the @devops agent.
- Use Case: Given an epic with five stories, run
aiox wave from-epic to generate a yolo-mode plan, dispatch ready stories in parallel batches, cascade-block dependents of failed stories, and produce a wave report for the devops merge step.
Quick Start
Ask the agent to execute wave WAVE-1 with your story files in yolo mode, for example: run wave-execute WAVE-1 with stories story-a.md and story-b.md in yolo mode with a dry run first.