spawn-agents

Spawn agents by calling the spawnAgents function with a taskId, loop type (loop3 or loop2), and agents array. The module validates agent types, injects task and iteration context, and supports dry-run mode for testing before execution.

Agent spawning validates that agent types are either loop3 or loop2, ensuring only supported loop implementations are orchestrated. Invalid types are rejected before container execution.

Yes, dry-run mode logs spawn commands and agent configurations without executing them, letting you verify task context injection and agent selection before live deployment.

Context injection attaches task and iteration metadata to spawned agents, ensuring each agent receives the required information for consensus collection and product owner decision flow in CFN Loop orchestration.

Agent spawning coordinates docker-based CFN Loop workflows with Redis-backed state management, managing parallel execution across Loop3 implementers and Loop2 validators with configurable timeouts and error handling.

Spawned agents execute tasks, report confidence levels, trigger gate checks, and aggregate consensus for GOAP-driven product owner decisions based on configured thresholds and adaptive agent selection.