No red flags found in any of the 11 categories — no credential harvesting, no data exfiltration, no curl-pipe-shell installer. Scanned against the SlowMist agent-security taxonomy, refreshed every 8 hours. Full audit →
by fubak · Agent Tool · ★ 83
Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h
🔒 Is ultraswarm safe to install? View the security audit →
ultraswarm Intelligent orchestration of external AI coding CLIs — complexity-routed models, isolated git worktrees, adversarial QA, Claude-only merge. orchestrates a swarm of external AI coding CLIs (codex, gemini, grok, agy, droid, opencode). It analyzes your task's complexity, decomposes it into atomic units, routes each to the best CLI and the right model tier within that CLI — fast/cheap models for simple work, powerful models for complex work, with automatic escalation when an attempt fails QA — runs them in isolated git worktrees, QA-reviews every diff, and merges only what passes. Dependent tasks run as chained waves, each re-based on the merged result of the wave before it. The orchestrator never writes feature code; it decomposes, reviews, judges, merges, and reports. Run it three ways (same orchestration core, identical behaviour): As a Claude Code skill — . Native Workflow engine with live progress and resume. → Claude Code installation As a Codex skill — . Codex plans and launches the standalone runner. → Codex installation As a standalone CLI — , hosted from Grok or any shell.
| Stars | 83 |
| Forks | 10 |
| Language | JavaScript |
| Category | Agent Tool |
| License | MIT |
| Quality Score | 65.3817370951434/100 |
| Open Issues | 1 |
| Last Updated | 2026-07-03 |
| Created | 2026-06-08 |
| Platforms | claude-code, cli, node |
| Est. Tokens | ~22k |
These tools work well together with ultraswarm for enhanced workflows:
Looking for a ultraswarm alternative? If you're comparing ultraswarm with other agent tool tools, these 6 projects are the closest alternatives on Agent Skills Hub — ranked by topic overlap, star count, and community traction.
Multi-agent orchestration system for Claude Code with parallel execution, automated quality gates, Board of Di
Local-first coordination for human and agent work: durable work, decisions, dispatches, evidence, and prompt-f
Define task-specific AI sub-agents in Markdown for any MCP-compatible tool.
Turn feature specs into merged PRs with a self-supervising swarm of coding agents — parallel execution, isolat
Automated loop driver, slash commands, council automation, MCP browser bridge, and portfolio governance for Cl
Kindly Web Search MCP Server: Web search + robust content retrieval for AI coding tools (Claude Code, Codex, C
Explore other popular agent tool tools:
ultraswarm is Multi-CLI agent swarm orchestrated by Claude Code: external AI CLIs code in isolated worktrees, Claude verifies and merges. It is categorized as a Agent Tool with 83 GitHub stars.
ultraswarm is primarily written in JavaScript. It covers topics such as ai-agents, claude-code, orchestration.
You can find installation instructions and usage details in the ultraswarm GitHub repository at github.com/fubak/ultraswarm. The project has 83 stars and 10 forks, indicating an active community.
ultraswarm is released under the MIT license, making it free to use and modify according to the license terms.
The top alternatives to ultraswarm on Agent Skills Hub include conductor-orchestrator-superpowers, maestro, sub-agents-mcp. Each offers a different approach to the same problem space — compare them side-by-side by stars, quality score, and community activity.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: