agent-swarm

Partition workloads across multiple agents and merge verified results.

5|3|Updated Feb 21, 2026
One-click install
npx skills add https://github.com/richfrem/agent-plugins-skills --skill agent-swarm-richfrem
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-swarm
Source: https://github.com/richfrem/agent-plugins-skills/tree/main/plugins/agent-loops/skills/agent-swarm
Command: npx skills add https://github.com/richfrem/agent-plugins-skills --skill agent-swarm-richfrem

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Work orchestration across multiple agents by partitioning large work into independent sub-tasks, routing execution modes, and validating/merging results to reduce overall completion time.

Core Features & Use Cases

  • Plan & Partition: Break work into independent tasks with clear boundaries.
  • Route: Choose Sequential Pipeline or Parallel Swarm based on task dependencies.
  • Dispatch: Create a per-task worktree and assign each to an agent (CLI, deterministic script, or human).
  • Execute: Run tasks in isolation with no cross-worktree interference.
  • Verify & Merge: Validate outputs against criteria and merge successful results.
  • Seal: Bundle merged artifacts for deployment or archival.
  • Retrospective: Evaluate partition strategy and parallelism effectiveness.

Quick Start

To begin, define a partitionable workload and run the swarm runner with your preferred engine and workers.

Frequently Asked Questions about agent-swarm

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

FAQPage Schema
How do I orchestrate parallel agents to complete large feature development faster?

Multi-agent orchestration partitions large workloads into independent sub-tasks, dispatching them across isolated worktrees to CLI agents, scripts, or humans, executing in parallel to reduce overall completion time.

What is the best way to partition bulk operations into independent sub-tasks for parallel execution?

Task partitioning breaks bulk operations into clearly bounded independent tasks, routing them as either a sequential pipeline or parallel swarm based on dependencies, then merging validated results.

Does multi-agent parallel execution work with deterministic scripts and CLI tools?

Parallel execution supports CLI agents, deterministic scripts, and human workers, creating per-task worktrees to run tasks in isolation without cross-worktree interference.

How do I verify and merge results from parallel agent workflows?

After parallel execution, the orchestration runner validates outputs against defined criteria, merges successful results, and bundles artifacts for deployment or archival.

Do I need Python 3.8+ to run multi-agent swarm orchestration?

Swarm orchestration requires Python 3.8+ and the standard library, using a swarm-like runner to partition, dispatch, and verify tasks across multiple agents.

When should I not use parallel multi-agent orchestration for task execution?

Avoid parallel swarm orchestration when workloads cannot be decomposed into independent sub-tasks, as dependent tasks require sequential pipeline routing instead of parallel dispatch.