How does it manage dependencies between agents?

It uses a task coordination strategy that minimizes deep dependency chains, preferring wide, shallow graphs and using blockedBy/blocks markers for integration boundaries.

What is the core rule of Agent Teams?

The 'Cardinal Rule' is one owner per file. To avoid concurrent modification issues, each file is assigned to a single agent at any given time.

What team compositions are available?

Available presets include Review, Debug, Feature, Fullstack, Research, Security, and Migration teams, each with specialized agent counts and roles.

Do I need to enable experimental features for this skill?

Yes, native parallel execution requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 in your environment, though the patterns also apply to manual parallel dispatch.

Agent Teams Workflow

Name: Agent Teams Workflow
Author: hjemmesidekongen

byhjemmesidekongen

0•

コラボレーションツール

Orchestrates parallel multi-agent teams for complex development tasks using specialized presets and file ownership rules.

Agent Teams provides a structured framework for parallel agent execution within Claude Code, allowing developers to scale their productivity by dispatching multiple specialized agents simultaneously. It includes pre-defined team compositions for common workflows—such as security audits, fullstack feature development, and code reviews—while enforcing a strict 'one owner per file' rule to prevent merge conflicts. By coordinating task dependencies and integration boundaries, this skill enables Claude to function as a collaborative workforce, handling large-scale migrations and multi-layered feature implementations through optimized vertical or horizontal slicing strategies.

主な機能

01Support for vertical, horizontal, and hybrid integration patterns

02Strict file ownership enforcement to prevent concurrent modification conflicts

030 GitHub stars

04Integration with Claude Code experimental Agent Teams API and manual dispatch

05Specialized team presets for Review, Feature, Security, and Debugging workflows

06Task coordination using wide, shallow dependency graphs

ユースケース

01Executing large-scale code migrations with a lead agent and multiple implementers working on conflict-free streams

02Performing comprehensive security audits using multiple agents to check OWASP compliance and dependencies

03Accelerating feature development by assigning frontend, backend, and testing tasks to parallel agents

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add hjemmesidekongen/ai agent-teams

For use in Claude.ai and ChatGPT