The Agency-Agents Repository: Your Complete AI Team in a Box

Why You Need This: The Ultimate AI Assistant Ecosystem

Imagine having instant access to 157 specialized experts who never sleep, never take breaks, and are always ready to tackle any task you throw at them. That’s exactly what the agency-agents repository delivers - a complete AI agency at your fingertips, ready to transform how you work, create, and solve problems.

What Is This For?

agency-agents is not just another collection of AI prompts. It’s a meticulously organized ecosystem of 157 specialized AI agents across 17 professional divisions, each designed to handle specific tasks with expert-level precision. Whether you’re:

  • Building a startup MVP and need a full engineering team
  • Launching a marketing campaign and require content creators, SEO specialists, and social media strategists
  • Developing enterprise software and need architects, testers, and documentation experts
  • Creating digital products and want UX designers, product managers, and quality assurance specialists
  • Exploring new technologies like spatial computing or game development

…this repository has the right specialist for every job.

Who Will Find This Helpful?

  1. Developers & Engineers: Get instant access to frontend wizards, backend architects, DevOps automators, and 23 other engineering specialists
  2. Marketers & Growth Hackers: Deploy 29 marketing agents including TikTok strategists, SEO specialists, content creators, and social media experts
  3. Designers & Creatives: Work with 8 design specialists from UI/UX architects to visual storytellers and brand guardians
  4. Product Managers & Entrepreneurs: Access product strategists, behavioral nudge engineers, and sprint prioritizers
  5. Business Professionals: Utilize sales coaches, account strategists, compliance auditors, and specialized industry experts
  6. Students & Researchers: Leverage academic anthropologists, historians, psychologists, and geographers
  7. AI Enthusiasts & Prompt Engineers: Study expertly crafted agent personalities and workflows for your own AI projects

The Core Value Proposition

What makes agency-agents truly revolutionary is that each agent isn’t just a generic prompt - it’s a complete personality with:

  • Specialized expertise in a specific domain
  • Proven workflows that deliver measurable results
  • Communication style tailored to their role
  • Memory and experience that improves over time
  • Integration capabilities with all major AI coding assistants

Quick Stats at a Glance

  • 157 AI Agents ready to deploy
  • 17 Specialized Divisions covering every professional domain
  • MIT Licensed - free to use, modify, and distribute
  • Multi-Tool Support - works with Claude Code, Cursor, Aider, Windsurf, Copilot, and more
  • Proven Workflows - includes real-world examples and implementation guides

Ready to Transform Your Workflow?

Whether you’re a solo developer needing a full team, a marketer launching campaigns, or an entrepreneur building products, agency-agents gives you the superpower of having 157 experts on call 24/7. The best part? They’re completely free, open-source, and ready to integrate into your favorite AI tools right now.


Repository Architecture

Directory Structure

agency-agents/
 .github/ # GitHub workflows and templates
 academic/ # 5 academic research agents
 design/ # 8 design and UX agents 
 engineering/ # 26 engineering and development agents
 examples/ # 6 workflow examples
 game-development/ # 5 game development specialists
 integrations/ # 1 integration guide
 marketing/ # 29 marketing and growth agents
 paid-media/ # 7 paid media specialists
 product/ # 5 product management agents
 project-management/ # 6 project management agents
 sales/ # 8 sales and business development agents
 scripts/ # 3 automation scripts
 spatial-computing/ # 6 spatial computing specialists
 specialized/ # 28 specialized industry agents
 strategy/ # 3 strategic planning agents
 support/ # 6 support and operations agents
 testing/ # 8 testing and quality assurance agents

Agent File Structure

Each agent follows a consistent Markdown format with frontmatter:

---
name: Frontend Developer
description: Expert frontend developer specializing in modern web technologies...
color: cyan
emoji: 
vibe: Builds responsive, accessible web apps with pixel-perfect precision.
---

Followed by detailed sections:

  • Identity & Memory: Role definition and personality traits
  • Core Mission: Primary responsibilities and capabilities
  • Critical Rules: Mandatory guidelines and constraints
  • Workflows: Step-by-step processes for common tasks
  • Deliverables: Concrete outputs and success metrics
  • Communication Style: How the agent interacts and presents information

Repository Statistics

Division Breakdown

Division Agent Count Percentage Key Specialties
Marketing 29 18.5% Social media, SEO, content creation, China market
Engineering 26 16.6% Frontend, backend, AI, DevOps, security
Specialized 28 17.8% Industry-specific, compliance, automation
Design 8 5.1% UI/UX, branding, visual storytelling
Sales 8 5.1% Outbound, discovery, deal strategy
Testing 8 5.1% QA, accessibility, performance testing
Paid Media 7 4.5% PPC, programmatic, social ads
Project Management 6 3.8% Agile, Jira, studio operations
Support 6 3.8% Analytics, finance, legal compliance
Spatial Computing 6 3.8% XR, VisionOS, immersive development
Examples 6 3.8% Workflow templates and demonstrations
Academic 5 3.2% Research, analysis, academic writing
Game Development 5 3.2% Game design, narrative, technical art
Product 5 3.2% Prioritization, feedback, behavioral design
Strategy 3 1.9% Executive briefs, nexus strategy
Integrations 1 0.6% Tool integration guides
Scripts 0 0.0% Automation scripts

Total: 157 agents across 17 divisions

Geographic and Market Focus

The repository shows strong internationalization, particularly with:

  • China Market: 12+ agents focused on Chinese platforms (WeChat, Douyin, Xiaohongshu, Bilibili, Baidu)
  • Global Coverage: Agents for major platforms worldwide (Twitter, LinkedIn, TikTok, Instagram, Reddit)
  • Cross-border: Specialists for Amazon, Shopee, Lazada, and other international e-commerce

Detailed Division Analysis

Engineering Division (26 Agents)

The largest technical division covering full-stack development:

Frontend & UI Development:

  • Frontend Developer (React/Vue/Angular)
  • UI Designer and UX Architect
  • WeChat Mini Program Developer
  • Feishu Integration Developer

Backend & Infrastructure:

  • Backend Architect (API design, databases)
  • DevOps Automator (CI/CD, cloud ops)
  • Database Optimizer (PostgreSQL/MySQL tuning)
  • SRE (Site Reliability Engineering)

Specialized Engineering:

  • AI Engineer (ML models, deployment)
  • Security Engineer (threat modeling, secure code)
  • Embedded Firmware Engineer (IoT, RTOS)
  • Solidity Smart Contract Engineer (EVM, DeFi)
  • Data Engineer (pipelines, warehousing)

Quality & Process:

  • Code Reviewer (PR reviews, quality gates)
  • Git Workflow Master (branching strategies)
  • Software Architect (system design, DDD)
  • Technical Writer (documentation)

Marketing Division (29 Agents)

The largest division with comprehensive digital marketing coverage:

Social Media Platforms:

  • Twitter Engager, TikTok Strategist, Instagram Curator
  • LinkedIn Content Creator, Reddit Community Builder
  • WeChat Official Account Manager, Weibo Strategist

Search & SEO:

  • SEO Specialist (technical SEO, content strategy)
  • Baidu SEO Specialist (China search optimization)
  • AI Citation Strategist (AEO/GEO for AI tools)

Content & Growth:

  • Content Creator (multi-platform content)
  • Growth Hacker (viral loops, experiments)
  • Book Co-Author (thought leadership)
  • Video Optimization Specialist (YouTube algorithm)

China Market Specialists:

  • Xiaohongshu Specialist (lifestyle content)
  • Bilibili Content Strategist (danmaku culture)
  • Douyin Strategist (short-video marketing)
  • Kuaishou Strategist (grassroots communities)
  • China E-Commerce Operator (Taobao, Tmall, Pinduoduo)

Specialized Division (28 Agents)

Industry-specific and niche expertise:

Compliance & Governance:

  • Compliance Auditor (regulatory compliance)
  • Healthcare Marketing Compliance (HIPAA, FDA)
  • Government Digital Presales Consultant
  • Blockchain Security Auditor

Automation & AI:

  • Agents Orchestrator (multi-agent coordination)
  • Automation Governance Architect
  • Data Consolidation Agent
  • Agentic Identity Trust

Business Operations:

  • Accounts Payable Agent (finance automation)
  • Corporate Training Designer (L&D programs)
  • Executive Summary Generator
  • Legal Compliance Checker

Design Division (8 Agents)

Creative and user experience specialists:

  • UI Designer (visual design, component libraries)
  • UX Researcher (user testing, behavior analysis)
  • UX Architect (technical architecture, CSS systems)
  • Brand Guardian (brand identity, consistency)
  • Visual Storyteller (visual narratives, multimedia)
  • Whimsy Injector (personality, delight, playful interactions)
  • Image Prompt Engineer (AI image generation)
  • Inclusive Visuals Specialist (representation, bias mitigation)

Technical Implementation

Agent File Format

Each agent file is a Markdown document with YAML frontmatter followed by structured content:

---
name: Agent Name
description: Comprehensive description
color: brand color
emoji: representative emoji
vibe: short tagline
---

# Agent Personality

## Your Identity & Memory
- **Role**: Specific domain expertise
- **Personality**: Character traits and approach
- **Memory**: What the agent remembers from past interactions
- **Experience**: Background and context

## Your Core Mission
### Primary Responsibility 1
- Detailed task breakdown
- Technical requirements
- Success criteria

### Primary Responsibility 2
- Additional capabilities
- Implementation guidelines
- Quality standards

## Critical Rules You Must Follow
1. Rule 1: Mandatory constraint
2. Rule 2: Ethical guideline
3. Rule 3: Technical requirement

## Your Workflows
### Standard Workflow
1. Step 1: Initial assessment
2. Step 2: Research phase
3. Step 3: Implementation
4. Step 4: Quality assurance
5. Step 5: Delivery and follow-up

## Your Deliverables
- **Code**: Production-ready implementations
- **Documentation**: Comprehensive guides and references
- **Processes**: Optimized workflows and best practices
- **Metrics**: Measurable outcomes and success indicators

## Your Communication Style
- **Tone**: Professional, enthusiastic, etc.
- **Format**: Structured responses, code examples
- **Feedback**: Constructive criticism approach
- **Collaboration**: Team interaction style

Integration Scripts

The repository includes automation scripts for tool integration:

  1. convert.sh: Generates integration files for all supported tools
  2. install.sh: Interactive installation that auto-detects installed tools
  3. lint-agents.sh: Validates agent file structure and consistency

Supported tools include:

  • Claude Code (primary target)
  • Cursor, Aider, Windsurf
  • Gemini CLI, OpenCode, Kimi Code
  • GitHub Copilot

Visual Representation

Main Structure Diagram

Agency-Agents Structure Diagram

Clean architectural diagram showing the repository structure

Key Features:

  • Structured Layout: Organized division hierarchy
  • Color Coding: Consistent with division color scheme
  • Grid Design: Clean, readable architecture visualization
  • Clear Connections: Well-defined relationships between divisions

Agent Workflow Diagram

Agent Workflow Diagram

Visualizing the complete workflow of an AI agent from inputs to outputs

Key Components:

  • Inputs: User requests, context, tools, memory, and constraints
  • Process: Agent’s specialized processing, decision-making, and execution
  • Outputs: Code, documentation, designs, strategies, and measurable outcomes
  • Feedback Loop: Continuous improvement through memory and experience
  • Clean Layout: Structured visualization for easy understanding

Integration Architecture Diagram

Integration Architecture

How agency-agents integrate with various AI tools and platforms

Integration Points:

  • AI Coding Assistants: Claude Code, Cursor, Aider, Windsurf, Copilot
  • Development Environments: VS Code, JetBrains, terminal integration
  • Memory Systems: MCP memory, context preservation, experience tracking
  • Tool Chains: Git, CI/CD, testing frameworks, deployment pipelines

Agent File Structure Diagram

Agent File Structure

Standardized file format and organization for all 157 agents

File Structure Components:

  • Identity & Memory: Agent personality, experience, communication style
  • Core Competencies: Primary skills and specialized knowledge areas
  • Workflow Examples: Step-by-step implementation guides
  • Success Metrics: Measurable outcomes and quality standards
  • Tool Integration: Supported tools and platform compatibility

Data Flow Diagram

Data Flow Diagram

Information transformation through the agent ecosystem

Data Flow Stages:

  1. Input Processing: User requirements → structured tasks
  2. Agent Selection: Task routing to appropriate specialists
  3. Execution Phase: Tool usage, code generation, content creation
  4. Output Generation: Deliverables with quality validation
  5. Feedback Integration: Results → memory → improvement

Usage and Integration

Quick Start Guide

Option 1: Claude Code Integration

# Copy agents to your Claude Code directory
cp -r agency-agents/* ~/.claude/agents/

# Activate any agent in Claude Code sessions:
# "Hey Claude, activate Frontend Developer mode and help me build a React component"

Option 2: Multi-Tool Integration

# Generate integration files for all supported tools
./scripts/convert.sh

# Install interactively (auto-detects installed tools)
./scripts/install.sh

# Or target specific tools directly
./scripts/install.sh --tool cursor
./scripts/install.sh --tool copilot
./scripts/install.sh --tool aider

Option 3: Reference Implementation

  • Browse agents and copy/adapt needed ones
  • Use as prompt templates for other AI systems
  • Study workflow patterns for team automation

Supported AI Tools

  1. Claude Code (Primary): Native agent activation
  2. Cursor: AI-powered code editor integration
  3. Aider: Command-line chat for code
  4. Windsurf: AI-native development environment
  5. GitHub Copilot: Prompt pattern integration
  6. Gemini CLI: Google’s AI command-line tool
  7. OpenCode: Open-source coding assistant
  8. Kimi Code: Chinese AI coding assistant

Case Studies

Case Study 1: Startup MVP Development

Scenario: A startup needs to build a minimum viable product quickly.

Agent Team:

  1. Product Manager: Defines requirements and prioritizes features
  2. Frontend Developer: Builds React components with TypeScript
  3. Backend Architect: Designs API and database schema
  4. DevOps Automator: Sets up CI/CD pipeline
  5. UI Designer: Creates user interface and design system

Workflow: Follows examples/workflow-startup-mvp.md template Result: Production-ready MVP in 2 weeks with automated deployment

Case Study 2: China Market Entry

Scenario: A Western company wants to enter the Chinese market.

Agent Team:

  1. China Market Localization Strategist: Overall GTM strategy
  2. Xiaohongshu Specialist: Lifestyle content creation
  3. Douyin Strategist: Short-video marketing campaigns
  4. WeChat Official Account Manager: Subscriber engagement
  5. Baidu SEO Specialist: Search engine optimization
  6. China E-Commerce Operator: Taobao/Tmall store setup

Workflow: Integrated cross-platform strategy with cultural adaptation Result: Successful market entry with localized brand presence


Key Takeaways

Strengths

  1. Comprehensive Coverage: 157 agents across virtually all business functions
  2. Practical Implementation: Production-ready workflows and deliverables
  3. International Perspective: Strong focus on global markets, especially China
  4. Technical Depth: Engineering agents cover full-stack development
  5. Community Focus: Active development and contribution encouragement

Unique Value Propositions

  1. Personality-Driven AI: Beyond functional prompts to character-based interaction
  2. Cross-Tool Compatibility: Works with multiple AI coding environments
  3. Industry Specialization: Deep expertise in niche domains
  4. Workflow Automation: Scripts for seamless integration and deployment
  5. Open Source Ecosystem: MIT license enables commercial use and modification

Practical Applications

  1. Development Teams: Accelerate coding with specialized AI assistants
  2. Marketing Agencies: Scale content creation across multiple platforms
  3. Startups: Access expert knowledge without hiring full teams
  4. Enterprises: Standardize processes with AI-driven best practices
  5. Education: Learn domain expertise through interactive AI mentorship

Conclusion

The agency-agents repository represents a significant advancement in practical AI application. By providing 157 specialized agent personalities across 17 business divisions, it offers a comprehensive toolkit for AI-assisted work in virtually any domain.

Key Innovations:

  • Personality-driven AI interactions beyond simple prompt engineering
  • Cross-platform compatibility with major AI coding assistants
  • Deep internationalization, particularly for Chinese markets
  • Production-ready workflows with measurable outcomes
  • Open-source ecosystem encouraging community collaboration

Whether you’re a developer, marketer, designer, or business professional, agency-agents gives you the power to work smarter, not harder, by leveraging the collective expertise of 157 specialized AI professionals - all available at your fingertips, 24/7.

Watch PyShine on YouTube

Contents