What does this skill do?

The Skill Agent Team Builder offers an interactive menu for exploring, composing, and deploying teams of AI agents in parallel on demand. It automatically detects available agents in your environment, groups them by domain, and runs simultaneous analyses on the same project, repository, or document, generating a unified report with agreements, conflicts, and prioritized next steps.

Agent Discovery
It lists all the agents available in your environment (users, plugins, and built-in agents) and automatically groups them by domain.
Multidisciplinary ad hoc teams
Combine experts from different fields (for example, Security + SEO + Architecture) to approach a task from multiple perspectives.
Parallel Analysis
Run multiple agents simultaneously on the same project, code repository, or document to speed up the delivery of results.
Summary of Recommendations
It compiles the interactions of all agents and generates a comprehensive report that highlights agreements, conflicts, and prioritized next steps.

Usage examples

🚀 Pre-launch review
Put together a team to review my Next.js e-commerce site before launch.
📈 Content Audit
I need an SEO specialist and a content strategist to analyze my blog and suggest ways to improve its search engine rankings.
🧩 Multidisciplinary team
Select domains 1 and 3 to audit my code repository in parallel.
📊 Expert Summary
Run Security Engineer and Software Architect on my project and provide me with a report detailing agreements and conflicts.

Features

Automatic Agent Detection Run `claude agents` and search in ./agents/**/*.md and ~/.claude/agents/**/*.md to list all available agents.
Grouping by domains Organize agents by subdirectory or file prefix, displaying the number of agents and their names by domain.
Parallel Deployment Launch parallel calls to the Agent tool by injecting the task description into each selected agent.
Flexible Recruitment Management It allows you to select by domain number or agent name, and prompts you to narrow down your selection if you choose more than 5 agents.
Consolidated Report with Summary Generate findings by agent and a final section listing agreements, conflicts, and prioritized next steps.

Frequently asked questions

The skill searches in ./agents/**/*.md and ~/.claude/agents/**/*.md. You can also use `claude agents` to list the agents recognized by the CLI.
If you select more than 5 agents, the skill will ask you to narrow down your selection to maintain the quality of the analysis and prevent overload.
The failure is noted inline (for example, "Security Engineer: failed — [reason]") and the report continues with the successful results from the rest of the team.
No. The skill uses the Agent tool built into Claude with the "general-purpose" subagent_type to dispatch agents in parallel.
Agent Team Builder — Parallel Deployment of Agents with Claude AI

¿Prefieres escuchar el contenido? Genera la narración de audio con un clic.