orchestrate

To run parallel AI agents with coordination, you need an orchestration workflow that spawns subagents concurrently while managing shared state. This Skill executes parallel agent workflows by spawning subagents through Gemini CLI and managing shared workflow state via MCP Memory.

MCP Memory is used in agent orchestration to share state, context, and coordination signals across multiple subagents. It allows the workflow to maintain a shared memory-driven state so parallel tasks stay synchronized and monitored without manual intervention.

To verify the results of a multi-agent workflow, you run a post-execution verification step that checks if outcomes match expected behavior. This Skill includes progress monitoring and runs a verification step after parallel agent execution to confirm reliability.

Yes, you can use Gemini CLI to automate concurrent development tasks by spawning coordinated subagents that run in parallel. This Skill targets development and automation scenarios, using Gemini CLI to execute multi-agent task breakdowns with deterministic coordination.

Yes, parallel agent orchestration requires following the workflow instructions defined in .agents/workflows/orchestrate.md. This file provides the repeatable operational procedure and step-by-step workflow that the Skill executes to manage deterministic coordination.

You should not use a multi-agent orchestration workflow for simple, sequential tasks that do not require task breakdowns or shared state. This Skill is designed for complex development and automation scenarios requiring multi-agent task breakdowns, progress monitoring, and post-run verification.