Swarm
- 6 repo stars
- Updated July 29, 2026
- GGPrompts/my-plugins
Run 10 Haiku agents in parallel on a question, then have them vote to pick the best answer.
About
An agent-building skill that fans a question out to ten Haiku agents thinking in parallel, then aggregates their outputs through a vote to choose the strongest answer. Developers use it when a single model pass is unreliable and a cheap ensemble-with-consensus would raise answer quality.
- Parallel multi-agent reasoning
- 10 Haiku workers
- Consensus by voting
- Better-answer selection
Swarm by the numbers
- Data as of Jul 30, 2026 (Skillselion catalog sync)
/plugin marketplace add GGPrompts/my-plugins/plugin install swarm@my-pluginsAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| repo stars | ★ 6 |
|---|---|
| Last updated | July 29, 2026 |
| Repository | GGPrompts/my-plugins ↗ |
What it does
Run 10 Haiku agents in parallel on a question, then have them vote to pick the best answer.
README.md
Swarm Plugin
Swarm consensus — 10 haiku agents think in parallel, then vote on the best answer.
Overview
Uses parallel haiku agents as independent reasoning units. Each agent thinks about the question independently, then all agents vote on the best answer. Democratic consensus emerges from diverse parallel reasoning.
Usage
/swarm What is the best approach to implement caching in this service?
How It Works
- THINK — 10 haiku agents reason independently in parallel
- Present — All answers displayed in a table
- VOTE — 10 agents evaluate all candidates and pick the best
- Report — Winner announced with vote distribution and consensus strength
Components
commands/swarm.md— The/swarmslash command (orchestrator)agents/haiku-swarm-skills.md— The haiku reasoning unit agent
Related skills
Setup Matt Pocock SkillsScaffold the per-repo configuration that Matt Pocock’s engineering agent skills rely on so they understand the issue tracker, triage labels, and domain documentation la495k196k
Lark AppsConnect Claude, Cursor or custom agents directly to Lark (Feishu) for messaging, document automation, approval workflows and enterprise data access.402k
CavemanSlash token usage by roughly 75% while keeping every technical detail intact when working with Claude Code, Cursor or similar agents.391k94.7k
Lark Skill MakerQuickly turn any Lark/Feishu OpenAPI call or multi-step workflow into a reusable agent skill with its own SKILL.md.386k16k
Running Claude Code Via Litellm CopilotRun Claude Code at a fraction of the cost by routing requests through LiteLLM to the GitHub Copilot Chat API.270k72
Codex PetGenerate a complete Codex Pet spritesheet and metadata from one reference image without needing an OpenAI key or Codex Pro.246k10