parallel-task-spark

Parallel task execution is automated by coordinating Sparky subagents in dependency-aware waves. Provide a plan file with task definitions, trigger the skill, and it parses tasks, enforces dependencies, launches subagents, and logs progress automatically.

Wave-based orchestration divides plan tasks into sequential execution waves based on dependency graphs. Only unblocked tasks run in each wave, ensuring dependent tasks wait for prerequisites to complete before launching Sparky subagents.

To start parallel plan execution, provide a plan file like plan.md containing task definitions and dependencies, then trigger the skill with /parallel-task-spark. The orchestrator parses the plan and begins wave-based subagent execution immediately.

Yes, parallel subagent orchestration is designed for plans with complex dependency chains. The system parses dependency relationships and only launches unblocked tasks, ensuring correct execution order while maximizing parallelism across independent tasks.

Yes, parallel task orchestration automatically updates plan docs with progress logs after each task completes. The system writes structured completion records back to the plan file, maintaining a persistent execution trail throughout all waves.

Wave-based parallel execution requires well-defined task definitions and explicit dependency mappings in the plan file. Plans with ambiguous dependencies or poorly structured task boundaries may cause subagents to block or execute out of intended order.