orchestrator

To orchestrate parallel execution across tmux sessions, load a plan with multiple specs and the orchestrator will launch and manage weavers concurrently. It reads plan metadata to schedule weavers based on dependencies, ensuring orderly progression across stages.

Dependency-aware scheduling reads spec dependencies within your plan metadata and enforces correct execution order for weavers. It determines which weavers can launch in parallel and which must wait, preventing out-of-order execution and automating plan-driven workflows.

Yes, you can use tmux to automate plan-driven workflows without manual coordination by loading a plan into the orchestrator. It manages weaver execution across sessions, tracks progress, and writes a final summary upon plan completion.

The best way to monitor weaver statuses and runtime state is to let the orchestrator write runtime state into .claude/vertical/plans. It tracks weaver statuses across tmux sessions and generates a final summary when the plan completes.

Yes, the orchestrator interfaces with the skill index to select weaver skills automatically. It reads your plan metadata to identify the required weavers and matches them with available skills before launching them in parallel via tmux.

You should avoid parallel weaver orchestration for plans lacking multiple specs or clear dependency definitions. The orchestrator is designed for complex plans with several interdependent specs; simpler, single-spec workflows do not require this automated coordination layer.