swarm

Dispatch table rows in parallel to subagents and merge results.

Updated May 29, 2026
One-click install
npx skills add https://github.com/patorn/langgraph-template --skill swarm-patorn
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: swarm
Source: https://github.com/patorn/langgraph-template/tree/main/.agents/skills/swarm
Command: npx skills add https://github.com/patorn/langgraph-template --skill swarm-patorn

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Dispatches many independent items in parallel: create a table, fan out to subagents, and aggregate results for faster outcomes.

Core Features & Use Cases

  • Parallel dispatch: execute multiple items across subagents concurrently with controlled batching.
  • Batch merging: aggregate per-row results back into a single table for downstream analysis.
  • Flexible sources: supports glob patterns, explicit file paths, or custom task rows.

Quick Start

Dispatch a set of rows by creating a table, run in parallel, and merge the results back into the table.

Frequently Asked Questions about swarm

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I dispatch tasks in parallel across table rows?

To dispatch tasks in parallel, you create a table of independent items, fan out the rows to subagents concurrently, and aggregate the results back into the table. Batching and concurrency controls manage the distribution.

What is the best way to process thousands of rows with subagents?

Processing thousands of rows with subagents is handled through controlled batching and parallel dispatch. The system scales from dozens to thousands of rows while enforcing structured output merging to maintain data integrity.

Can I use glob patterns or explicit file paths as task sources?

You can use flexible sources for task generation including glob patterns, explicit file paths, or custom task rows. The system fans out these inputs across subagents concurrently for processing.

How does batch merging handle errors during parallel dispatch?

Batch merging aggregates per-row results back into a single table while enforcing error handling. This structured output merging maintains data integrity even when dispatching tasks concurrently across subagents.

When do I need parallel dispatch for subagent task processing?

You need parallel dispatch when you have many independent items to process and want faster outcomes. By fanning out to subagents concurrently with controlled batching, you accelerate processing and aggregate results efficiently.