productionos-auto-swarm

To decompose complex Codex tasks for parallel execution, this Skill automatically breaks down the required task into slices and spawns a parallel subagent swarm to tackle them concurrently. It then orchestrates waves of execution to converge on a high-quality solution.

Parallel subagent swarm orchestration is a workflow mechanism that spawns multiple subagents to tackle task slices concurrently. It works by applying task decomposition, tracking a coverage map, and guiding iterative waves toward convergence with guardrails that stop execution when coverage is sufficient.

You can run a parallel swarm for a large code review task by providing the required task input alongside optional parameters like mode, depth, swarm_size, and iterations. The orchestrator automatically slices the review into parallel subagent waves and synthesizes the results per wave.

Yes, you can control the swarm size and depth of parallel subagent execution by adjusting the optional swarm_size and depth inputs. These parameters dictate how many subagents run concurrently and how extensively the task is decomposed across iterative waves.

You should use parallel swarm orchestration instead of sequential processing when facing large, complex analysis or code-review tasks that require broad coverage and accelerated delivery. It increases speed and depth of analysis by running concurrent subagents rather than handling tasks sequentially.

Yes, parallel swarm orchestration works without external dependencies, as it operates entirely within the Codex workflow environment. You only need to provide the required task input and optional parameters like mode, depth, swarm_size, and iterations to begin execution.