BrowseFull catalogOutcomesSolve a specific problemRolesStack by teamTrustFilter by risk tier
← Back to the Claude Observatory

swarms

Skill Productivity Usable
Works inClaude Code
Usable Scanned — metadata only

Getting out-of-the-box agent teams to converge takes real configuration effort.

Swarm is a Claude Code plugin that assembles a team of five to eight agents — engineers plus customer and business voices — that research the problem, debate the approach, build it, and review each other's work until it is a reviewed, ship-ready PR. You approve the team and plan, then it runs unattended, even across days. Also runs triage, writing, and general-purpose teams.

84 starsMIT (commercial OK)FreeQuick setup
Usable rating — This tool is functional but has notable gaps. Review the evaluation notes below before deploying.

Getting out-of-the-box agent teams to converge takes real configuration effort. Swarm assembles the specialist team for you, has work reviewed by a different agent than its author, runs unattended for hours or days, survives interrupted turns, and avoids the rate-limit errors parallel teams hit — at a fraction of the tokens.

Engineering teams executing multi-stage data pipelines or batch processing tasks requiring coordinated Claude API calls.

Claude Code Claude Cowork Claude Chat

https://github.com/DheerG/swarms

By DheerG

How to Get It

Option 1: Claude Desktop App (Code Mode)Click the + button next to the prompt box → PluginsAdd plugin. Search and click Install. Skills work in Claude Code only.
Option 2: Paste into Claude CodeCopy the command below and paste it into your conversation. Claude will install it.
Instructions to paste into Claude
claude plugin marketplace add DheerG/swarms
claude plugin install swarm@swarms --scope project

Enable agent teams in ~/.claude/settings.json: {"env": {"CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1"}} (or let /swarm:code enable it on first run), then restart Claude Code and run /swarm:code.
First thing to try

After installing, paste this into Claude:

Help me ship a reviewed, ready-to-merge PR from a one- or two-sentence feature description
PrerequisitesClaude Code CLI v2.1.178 or newer; agent teams enabled via CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 in ~/.claude/settings.json.CostFree

Trust Signals Auto-scanned

Stars84Last updated2026-07-23LicenseMIT (OK for commercial use)Known CVEsNone foundSources: GitHub Advisory Database + OSV.dev · Scanned 2026-08-12 · scanner vattempted-no-data

Community Pulse Emerging

Discussed on Hacker News

2 mentions across 1 sources

Reviewer notes

Auto-scanned review. These are observations, not a security certification.

Auto-assessment from April sweep — baseline scores pending hands-on review.

2026-04-18: Approved in April sweep: New and small but has a clear description and recent commits — worth tracking.

Things to check

  • Swarms abstracts agent coordination but adds latency for sequential pipelines. Token costs scale with agent count and task complexity; no built-in cost optimization. Quality depends on clear task decomposition and agent role definition.

How to evaluate tools before deploying →

Data shown here comes from public APIs and automated scanning. Reviewer notes reflect one person's experience. This is not a security certification or legal recommendation. Always evaluate tools according to your own organization's policies.

Evaluation

Ease of Use
3/5
Versatility
3/5
Reliability
3/5
Security
3/5
Overall score3.00 / 5.00 UsableEvaluatedApr 2026

← Back to the Claude Observatory

Rolling Claude out in your org? Let's talk.

Start a conversation →