/using-orchestration

name: using-orchestration
description: Use when user has complex multi-agent workflows, needs to coordinate sequential or parallel agent execution, wants workflow visualization and control, or mentions automating repetitive multi-agent processes - guides discovery and usage of the orchestration system

Using the Orchestration System

When to Suggest

Proactively suggest `/orchestrate` when user describes:

**Clear Triggers:**

• Multiple agents needed in sequence: "explore then review then implement"
• Parallel operations: "run tests and lint at the same time"
• Conditional workflows: "if tests pass, deploy; if failed, fix and retry"
• Retry logic: "keep trying until it works"
• Workflow automation: "I do this same process often"
• Multiple checkpoints: "I want to review before deploying"

**Specific Phrases:**

• "coordinate multiple agents"
• "workflow", "pipeline", "orchestrate"
• "run these in parallel"
• "if X then Y"
• "retry until success"
• "automate this process"

How to Introduce

When triggers detected, suggest proactively:

This workflow would benefit from orchestration. I can use `/orchestrate` to:
- Visualize the workflow graph
- Execute agents in parallel where possible
- Handle retries and conditionals
- Provide interactive control at checkpoints

Would you like me to use `/orchestrate [workflow-syntax]` or create a reusable template?

**First-time setup reminder:**

If user mentions custom agents (like `expert-code-implementer`, `code-optimizer`, etc.) that aren't in the orchestration namespace:

I notice you have custom agents in your environment.
Would you like to import them to orchestration first?

Run: /orchestration:init

This will make your custom agents available in workflows as:
- orchestration:expert-code-implementer
- orchestration:code-optimizer
- etc.

Quick Syntax Reference

Show relevant syntax based on user's needs:

**Sequential:** `explore:"task" -> review -> implement` **Parallel:** `[test || lint || security]` **Conditional:** `test (if passed)~> deploy` **Retry:** `@try -> fix -> test (if failed)~> @try`

When NOT to Suggest

• Single agent task (no coordination needed)
• Simple sequential tasks (<3 steps, no conditionals)
• User explicitly requests manual coordination
• Task already in progress without orchestration

Integration Pattern

When user agrees to orchestration: 1. Parse their requirements into workflow syntax 2. Invoke `/orchestrate [workflow-syntax]` 3. Let the orchestration system handle execution 4. Don't manually coordinate agents - let orchestrator do it