Awesome Claude Code Subagents

What it does

• Positions itself as the definitive collection of Claude Code subagents — specialized AI assistants for specific development tasks
• Coverage spans backend, frontend, and mobile language specialists, DevOps/SRE roles, QA and security reviewers, and documentation writers
• Documents the subagent mechanics end to end: create via the /agents manager (project or global), let Claude draft an initial version, refine purpose and activation triggers, configure tool access (empty inherits all), edit the system prompt, and engage automatically or by explicit request
• Standardized template with name, description (when to invoke), tools, and model fields; storage split between .claude/agents/ (project, higher precedence) and ~/.claude/agents/ (global, lower), project overriding global on name conflicts

When to use it

Pick a specialist instead of writing one from scratch — browse by role, copy the definition into .claude/agents/, and adjust the description and tool list to your repo before first use.

Getting started

Install as a Claude Code plugin (the README's recommended path): claude plugin marketplace add VoltAgent/awesome-claude-code-subagents, then claude plugin install <plugin-name>. Manual alternative — definitions live in your repo, storage layout:

{
  "project": ".claude/agents/ (current project only, wins name conflicts)",
  "global": "~/.claude/agents/ (all projects, lower precedence)"
}

Create the first one interactively with /agents and let Claude draft it.

Things to know

The repo explicitly does not accept PRs whose primary purpose is promoting a product or company — subagents must stay vendor-neutral; sponsorship is the supported way to surface a project.