A tool for launching a swarm, meaning several coding agents working on one request in parallel. It starts the agents through the Agent Toolkit and coordinates their work through handoff files and isolated workspaces.

Use it to start pair, team, or full parallel agent runs for a shared development task.

It helps divide work among multiple agents while keeping their outputs separated and shareable. It is useful when a task benefits from parallel roles or team-style collaboration.

swarm — Skill for Claude Code by ulises-jeremias/agent-toolkit

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add ulises-jeremias/agent-toolkit --skill swarm

Clone the repo

git clone --depth 1 https://github.com/ulises-jeremias/agent-toolkit

Made for: Claude Code.

Or install agent-toolkit-complete, the plugin that ships this one along with the rest of its 106 skills.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/ulises-jeremias/agent-toolkit/swarm/github.svg)](https://agentmods.dev/skills/ulises-jeremias/agent-toolkit/swarm)

Your own site

<a href="https://agentmods.dev/skills/ulises-jeremias/agent-toolkit/swarm"><img src="https://agentmods.dev/badge/skills/ulises-jeremias/agent-toolkit/swarm/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/ulises-jeremias/agent-toolkit/swarm"><img src="https://agentmods.dev/badge/skills/ulises-jeremias/agent-toolkit/swarm.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 31 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,248 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits

NVIDIA SkillSpector pass 7 Sept 2026

How audits are shown

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00031	$0.01248
Opus 5.5	$0.00012	$0.00499
Sonnet 5.5	$0.00006	$0.00250
Haiku 4.5	$0.00003	$0.00125

Measured 1mo ago against content hash d55d5a4a7756, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

swarm scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/agent-toolkit-complete/.github/skills/swarm/SKILL.md · 109 lines

How it starts

The opening of the file, as written. The whole thing — 109 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Swarm

Launch a production Agent Toolkit swarm from any natural language request via the agent-toolkit swarm CLI. Herdr is the default UI for an excellent interactive experience; tmux is the isolated fallback.

When to use

User says "launch a swarm", "start a swarm", "swarm this task", or describes parallel/team work that benefits from multiple agents.
Any task that mentions swarm, pair/team/full, Herdr, tmux, or worktree-per-writer.
User wants to reuse an existing swarm run for a new request (same windows, new handoff).

Prerequisites

agent-toolkit installed and on PATH (agent-toolkit swarm doctor should pass).
For Herdr: herdr CLI installed (herdr --version); Herdr daemon running.
For tmux: tmux installed; isolated socket is agent-toolkit-swarm-<run-id>.
Run inside a Git repo (or pass --workspace /path/to/repo).
Optional: ~/.config/agent-toolkit/swarm.yaml for model-profile mapping (economy → opencode/mimo-v2.5-free, balanced → provider default).

Workflow

1. Inspect intent

Extract from the user prompt:

Task: free-form objective (if missing, start swarm ready for interactive input).
Recipe: pair (2 roles: implementer→reviewer, fast), team (implementer→reviewer→integrator), full (large, 5+ roles). Default pair unless task needs more.
Runner: opencode (default, free models) or claude (Anthropic). Respect --runner if user names it.
Model profile: economy (opencode/mimo-v2.5-free) or balanced (provider default). Default balanced.
UI: herdr (default) or tmux (isolated). Respect --ui if given.

2. Check health

agent-toolkit swarm doctor
agent-toolkit swarm recipes

If doctor fails, report diagnostics and fix herdr/tmux/runner availability before launching.

3. Launch

Always use the CLI; never fabricate state.json or backend state.

# Herdr (preferred) — eager windows, shell-aware (detects $SHELL/zsh), attach by default
agent-toolkit swarm start --recipe pair --ui herdr --runner opencode --model-profile balanced --attach "task"

# Tmux (fallback) — isolated socket
agent-toolkit swarm start --recipe team --ui tmux --runner claude --model-profile economy --attach "task"

# Headless / no prompt — starts swarm ready, no initial task
agent-toolkit swarm start --recipe pair --ui herdr --runner opencode --model-profile balanced --attach

# Reuse same run/windows for a new request
agent-toolkit swarm handoff create --type artifact --from implementer --to reviewer --artifact "artifacts/new-task.md" --run-id <run-id>

Read the full file on GitHub · 109 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

1mo ago First seen · 109 lines · 31 tokens per session scan A d55d5a4a7756

Subscribe to this mod's changes

swarm is a skill published in the GitHub repository ulises-jeremias/agent-toolkit (19 stars, last pushed yesterday), licensed MIT. It adds 31 tokens to every session and 1,248 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-04.

Building agents Claude Code agent-toolkit ai-agents ai-tools cursor github-copilot

verify

Fact-check claims encountered during reading — dates, names, events, citations. Use when encountering historical facts or disputed claims.

yha9806/academic-writing-toolkit · 26 tokens

skillnote

Self-hosted skill registry for OpenClaw. Stores procedures your team writes (name, description, body), syncs them to disk before each task, and collects which-helped/which-failed signals from the agent so the registry improves over time.

luna-prompts/skillnote · 54 tokens

pr-alignment-loop

Iterate a PR to its final form by running two opposing reviewer droids (reviewer-robustness and reviewer-minimalist) in a bounded back-and-forth loop. The main orchestrator synthesizes their feedback, makes edits, runs tests, and stops when both approve, on stagnation, or after 3 rounds. Use when the user asks to…

nikships/skills-registry · 122 tokens

publish-website

Publish a website to Porkbun Secure Static Hosting. Use when the user asks you to build, create, update or put live a website, landing page or placeholder page for a domain they have at Porkbun, mentions Porkbun hosting or static hosting, or asks how to upload site files to Porkbun.

oborseth/Porkbun-MCP · 65 tokens

create-rule

Use when creating, writing, or updating a Claude Code rule file (.claude/rules/.md or /.claude/rules/.md), or when asking about rule format, path scoping, or rule best practices.

paultyng/skill-issue · 48 tokens

platform-setup

Set up, verify, and troubleshoot Katalon True Platform/TestOps MCP connectivity for any AI coding agent. Use when you need to install or configure the Katalon MCP server, connect to a Katalon platform account, verify available Katalon MCP tools, diagnose missing authentication or repository access, confirm…

katalon-labs/true-skills · 110 tokens