@justloveabit: 100 ready-to-use sub-agents that upgrade Claude Code from an 'assistant' to a 'complete AI company'. Most people still use Claude Code for single-agent chat. The real experts are already using 100+ specialized sub-agents to build their own AI engineering teams. Awesome repo...

X AI KOLs Timeline Tools
claude-code subagents multi-agent ai-agents open-source awesome-list

Summary

Introduces an Awesome repository containing over 100 sub-agents that upgrade Claude Code from a single-agent assistant to a multi-agent collaborative AI development platform, and provides a quick-start guide and top sub-agent recommendations.

100 ready-to-use sub-agents that upgrade Claude Code from an 'assistant' to a 'complete AI company'. Most people still use Claude Code for single-agent chat. The real experts are already using 100+ specialized sub-agents to build their own AI engineering teams. Awesome repository: Awesome Claude Code Subagents It contains over 100 ready-to-use sub-agents that upgrade Claude Code from an 'assistant' to a 'complete AI company'. Why are sub-agents so appealing? Single-agent is prone to hallucinations and context loss Sub-agents have clear division of labor + dedicated memory + MCP tool invocation One prompt = complete pipeline (plan → execute → review → deploy) After using them, my personal development efficiency has directly increased by more than 3 times. How to get started quickly? (4 steps) 1. Install the latest Claude Code Desktop / CLI 2. Fork the Awesome Claude Code Subagents repository 3. Put the corresponding sub-agent folder into your project's .claude/agents/ directory 4. Directly invoke in chat using @ agent_name. Top 10 Sub-agent recommendations (with invocation methods): 1. Architect Agent Responsible for system design, architecture diagrams, module decomposition Invocation: @ architect [project requirements] 2. Code Reviewer Agent Professional code review, security scanning, performance optimization Invocation: @ review [PR link or code block] 3. Test Engineer Agent Automatically generates unit tests, boundary cases, integration tests Invocation: @ tester [function or module] 4. Docs Writer Agent Generates README, API documentation, user manuals Invocation: @ docs [code path] 5. Bug Hunter Agent Specifically finds bugs, reproduces them, proposes fixes Invocation: @ hunter [error log] 6. Refactor Master Agent Code refactoring, modernization, performance improvement Invocation: @ refactor [module] 7. PM Agent (Product Manager) Requirement breakdown, user stories, priority sorting Invocation: @ pm [product idea] 8. DevOps Agent CI/CD configuration, Docker, deployment scripts Invocation: @ devops [project] 9. Research Agent Check latest technologies, competitive analysis, paper summaries Invocation: @ research [technical keywords] 10. Security Agent Security audit, vulnerability scanning, compliance check Invocation: @ sec [code] These sub-agents are not simple prompts; they come with: Dedicated CLAUDE.md / AGENTS.md Independent memory layer Dedicated MCP tool permissions Fixed workflow templates The repository has all ready-made configurations, one-click import. Repository highlights: Continuously updated (currently 100+) Community contribution mechanism Supports Claude 4 / Opus / Sonnet full series Compatible with Cursor / Windsurf / OpenCode and other tools Advanced usage: Use /goal + sub-agents combination to let them form temporary teams by themselves. For example: @ pm @ architect @ coder to build an AI Agent market analysis tool Claude will automatically schedule multiple sub-agents to collaborate. In 2026, the single-agent era is over. Sub-agents are your 'AI employees' — the more specialized they are, the closer you are to a one-person company. Which sub-agents are you currently using? Turn Claude Code into your own AI empire. #ClaudeCode #Subagents #AIAgents #MultiAgent #OnePersonCompany #AwesomeClaudeCode GitHub address: placed in the comments
Original Article
View Cached Full Text

Cached at: 05/25/26, 12:38 AM

100 out-of-the-box subagents that upgrade Claude Code from an “assistant” to a full “AI company”.

Most people still use Claude Code as a “single-agent chat”. But true experts are already using 100+ specialized subagents to build their own AI engineering team.

Awesome repository: Awesome Claude Code Subagents
It includes 100+ ready-to-use subagents that directly upgrade Claude Code from an “assistant” to a “complete AI company”.

Why are subagents so powerful?

  • Single agents are prone to hallucinations and context loss
  • Subagents have clear division of labor + dedicated memory + MCP tool calls
  • One prompt = complete pipeline (plan → execute → review → deploy)
  • After using them, my personal development efficiency has increased by more than 3x.

How to get started quickly? (4 steps)

  1. Install the latest Claude Code Desktop / CLI
  2. Fork the Awesome Claude Code Subagents repository
  3. Drop the corresponding subagent folder into your project’s .claude/agents/ directory
  4. Invoke it directly in chat with @agent_name.

Top 10 subagent recommendations (with invocation methods):

  1. Architect Agent
    Responsible for system design, architecture diagrams, module decomposition
    Invoke: @architect [project requirements]

  2. Code Reviewer Agent
    Professional code review, security scanning, performance optimization
    Invoke: @review [PR link or code block]

  3. Test Engineer Agent
    Automatically generate unit tests, edge cases, integration tests
    Invoke: @tester [function or module]

  4. Docs Writer Agent
    Generate README, API documentation, user manuals
    Invoke: @docs [code path]

  5. Bug Hunter Agent
    Specifically finds bugs, reproduces them, and suggests fixes
    Invoke: @hunter [error log]

  6. Refactor Master Agent
    Code refactoring, modernization, performance improvement
    Invoke: @refactor [module]

  7. PM Agent (Product Manager)
    Requirement breakdown, user stories, priority sorting
    Invoke: @pm [product idea]

  8. DevOps Agent
    CI/CD configuration, Docker, deployment scripts
    Invoke: @devops [project]

  9. Research Agent
    Look up latest technologies, competitive analysis, paper summarization
    Invoke: @research [technology keyword]

  10. Security Agent
    Security audit, vulnerability scanning, compliance checks
    Invoke: @sec [code]

These subagents are not simple prompts — they come with:

  • Dedicated CLAUDE.md / AGENTS.md
  • Independent memory layers
  • Dedicated MCP tool permissions
  • Fixed workflow templates
    All are pre-configured in the repo, one-click import.

Repository highlights:

  • Continuously updated (currently 100+)
  • Community contribution mechanism
  • Supports Claude 4 / Opus / Sonnet full series
  • Compatible with Cursor / Windsurf / OpenCode and other tools

Advanced usage:
Combine /goal with subagents to let them form a temporary team.
For example:
@pm @architect @coder build an AI agent market analysis tool
Claude will automatically schedule multiple subagents to collaborate.

In 2026, the era of single agents is over.
Subagents are your “AI employees” — the more and more specialized they are, the closer you get to a one-person company.

Which subagents are you currently using?
Turn Claude Code into your own AI empire.

#ClaudeCode #Subagents #AIAgents #MultiAgent #OnePersonCompany #AwesomeClaudeCode

GitHub link: in the comments

Similar Articles

@FakeMaidenMaker: If you really want to understand how the underlying of AI Agents like Claude Code is built, this open-source project writes one from scratch for you to see. GitHub has garnered 66.5K Stars, and also made it to the Trendshift hot list. The intelligence of an Agent comes from the model itself; what you can do is not "building intelligence"...

X AI KOLs Timeline

An open-source project teaches you to build a simplified version of Claude Code from scratch, thoroughly explaining the harness engineering of AI Agents. It has received 66.5K Stars.

@VincentLogic: Found an explosive open-source project with nearly 100k stars! agency-agents — a 'digital agency' that equips you with 140 AI employees. The nearly 100k stars on GitHub are no bluff! This is not just a single AI tool, but a complete set of multi-agent roles with job descriptions: 14 departments fully covered: …

X AI KOLs Timeline

Introducing agency-agents, an open-source project with nearly 100k stars on GitHub. It provides 140 specialized AI employee roles covering 14 departments, which can be directly integrated into AI coding assistants like Claude Code.

@AYi_AInotes: Damn, this GitHub project literally gives you an entire AI company — save and take it! 100k GitHub stars, called the fastest-growing AI project of 2026, 146 professional AI experts, 12 complete departments. One command, all packed into your Claude Code, from…

X AI KOLs Timeline

This GitHub project integrates an entire AI company, including 146 AI experts and 12 departments, and can be used in Claude Code with a single command, covering frontend development, security auditing, and more.

@NFTCPS: Guys, I have to recommend this project to you! Link: https://github.com/msitarzewski/agency-agents… There's an agent role library called agency-agents on GitHub with 81k stars—it's not playing around, it's genuinely exploding. 1…

X AI KOLs Timeline

Recommends the agency-agents project on GitHub, which includes over 140 expert-level AI agent roles (such as CEO, lawyer, programmer, etc.). It can be used with tools like Claude Code and Cursor, essentially forming a virtual AI team—free and open source.