SkillsApr 9, 2026·3 min read

Awesome Claude Code Subagents — 130+ Specialized Agents

Install 130+ specialized Claude Code subagents across 10 categories: core dev, language experts, infra, security, data/AI, DevEx, and business. Plugin-based with 16.7K GitHub stars.

TL;DR
A curated collection of 130+ Claude Code subagents covering core dev, security, data, infra, and more via plugin marketplace.
§01

What it is

Awesome Claude Code Subagents is a curated collection of over 130 specialized subagents for Claude Code, organized across 10 categories: core development, language-specific experts, infrastructure, security, data and AI, developer experience, and business operations. Each subagent is a focused expert that Claude Code can delegate tasks to for better results in specific domains.

This collection is designed for Claude Code users who want to extend their agent's capabilities with domain-specific expertise without building custom skills from scratch.

§02

How it saves time or tokens

General-purpose prompting produces adequate results, but specialized subagents produce expert-level output in their domain. A security-focused subagent knows common vulnerability patterns, OWASP guidelines, and secure coding practices without you providing that context each time. By delegating to the right subagent, you get better results in fewer iterations, saving both time and token usage.

§03

How to use

  1. Add the subagent collection via the plugin marketplace:
/plugin marketplace add https://github.com/VoltAgent/awesome-claude-code-subagents
  1. Install the categories you need:
/plugin install voltagent-core-dev
/plugin install voltagent-lang
/plugin install voltagent-infra
/plugin install voltagent-security
  1. Claude Code automatically delegates to the appropriate subagent based on the task context. You can also invoke subagents directly by name.
§04

Example

Available subagent categories and their specializations:

Category          Count   Examples
-----------       -----   --------
Core Dev            15    code-reviewer, refactorer, test-writer
Language Experts    20    python-expert, rust-expert, go-expert
Infra               12    docker-expert, k8s-expert, terraform-expert
Security            10    security-auditor, dependency-scanner
Data & AI           15    ml-pipeline, data-engineer, prompt-engineer
DevEx               10    git-workflow, ci-cd-expert, docs-writer
Business            8     api-designer, schema-architect
Frontend            12    react-expert, css-expert, a11y-auditor
Backend             15    db-optimizer, cache-expert, queue-designer
Mobile              8     ios-expert, android-expert, flutter-expert
§05

Related on TokRepo

§06

Common pitfalls

  • Installing all 130+ subagents at once. This adds significant context overhead. Install only the categories relevant to your current project and add more as needed.
  • Expecting subagents to work independently of your project context. Subagents still need access to your codebase and project files to produce relevant results. They add expertise, not project knowledge.
  • Not updating the plugin after initial installation. The collection receives regular updates with new subagents and improvements to existing ones. Run plugin update periodically.
  • Starting with an overly complex configuration instead of defaults. Begin with the minimal setup, verify it works, then customize incrementally. This approach catches configuration errors early and keeps troubleshooting straightforward.

Frequently Asked Questions

How do subagents differ from Claude Code skills?+

Skills provide instructions and patterns that Claude Code follows. Subagents are more specialized: they are focused expert personas that Claude Code delegates specific tasks to. A subagent has deep domain knowledge in a narrow area, while skills are broader workflow instructions.

Can I create my own subagents?+

Yes. The collection follows a standard format. You can create custom subagents by following the template structure and adding them to your project's plugin configuration. The README in the repository documents the subagent authoring format.

Do subagents consume additional tokens?+

Subagents add context to the conversation, which does consume tokens. However, they typically reduce total token usage because they produce better results in fewer iterations compared to generic prompting.

Which subagents are most popular?+

The core-dev category (code-reviewer, refactorer, test-writer) and security (security-auditor) are the most widely installed. Language-specific experts (python-expert, rust-expert) are also popular for teams working in specific technology stacks.

Do subagents work with other AI coding tools?+

This collection is specifically designed for Claude Code's plugin system. Other AI coding tools have their own extension mechanisms. The subagent prompts could be adapted for other tools, but the plugin installation and delegation system is Claude Code specific.

Citations (3)
🙏

Source & Thanks

Created by VoltAgent.

awesome-claude-code-subagents — ⭐ 16,700+

Thanks to VoltAgent for building the most comprehensive subagent collection for Claude Code.

Discussion

Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.