What problem does it solve? Complex tasks requiring multiple specialized agents often lose progress to context limits, lack coordination between parallel workstreams, and have no recovery mechanism when failures occur. This Skill provides structured multi-agent workflow management with persistent state, sync barriers for cross-pipeline information exchange, and checkpoint-based recovery. ## Core Features & Use Cases - Cross-Pollinated Pipelines: Run parallel agent pipelines (e.g., problem-solving and systems engineering) with bidirectional findings exchange at sync barriers. - Persistent State Management: Track workflow execution in ORCHESTRATION.yaml as a machine-readable single source of truth, surviving context compaction and session boundaries. - Checkpoint Recovery: Create recovery points after phases and barriers so long-running workflows can resume from any checkpoint. - Adversarial Quality Gates: Enforce creator-critic-revision cycles with a 0.92 quality threshold at every phase transition and barrier. - Use Case: Coordinate a design review where a research pipeline and a requirements pipeline run in parallel, exchange findings at two sync barriers, and produce a final synthesis combining both tracks. ## Quick Start Ask the AI to initialize an orchestration workflow for your project with a cross-pollinated pipeline using two skill tracks and sync barriers after each phase.
