claude-code-agents-team-nation-of-elites
advisely/claude-code-agents-team-nation-of-elites
Instalação
npx convez-ai advisely/claude-code-agents-team-nation-of-elites1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
README
A tradução para o português deste README ainda está na fila.
Claude Code Agents - Team Nation of Elites
An orchestrated AI workforce that functions like a real-world company.
Supercharge your work with a complete, role-based AI team that takes projects from concept to production — and from lead to revenue.
Created by Yassine Boumiza
Quick Start | Agent Roster | Skills | Docs | Contributing
The Problem & The Solution
A single, generalist AI, no matter how powerful, lacks the specialized expertise and structured workflow of a real-world company. Complex projects require a team of specialists -- architects, managers, developers, testers, and security experts -- all working in concert.
The "Nation of Elites" is the solution. It's not just a collection of agents; it's a fully-realized, virtual IT organization. Each agent has a specific role, a clear mission, and defined responsibilities, mirroring the structure of a high-performing tech company.
Platform Compatibility
Agents are platform-independent .md files -- they work identically on Windows, Linux, macOS, and WSL2.
| Interface | Supported | Config Path |
|---|---|---|
| Claude Code CLI (terminal) | Yes | ~/.claude/agents/ |
| VS Code + Claude Code extension | Yes | ~/.claude/agents/ |
| JetBrains + Claude Code extension | Yes | ~/.claude/agents/ |
Quick Start
Get your AI workforce operational in under one minute.
Prerequisites
| Platform | Requirement |
|---|---|
| Windows | Git for Windows |
| Linux / WSL2 | git, rsync |
| macOS | git (included with Xcode CLI tools) |
Step 1: Install Claude Code
# npm (if Node.js is installed)
npm install -g @anthropic-ai/claude-code
# Or use the official installer
curl -fsSL https://claude.ai/install.sh | sh
# PowerShell (recommended)
irm https://claude.ai/install.ps1 | iex
# Or via WinGet
winget install Anthropic.ClaudeCode
Git for Windows is required. If using a portable Git, set:
$env:CLAUDE_CODE_GIT_BASH_PATH = "C:\Program Files\Git\bin\bash.exe"
Verify with claude doctor.
Step 2: Deploy the Agents
Automated Deployment (Recommended)
git clone https://github.com/advisely/claude-code-agents-team-nation-of-elites.git
cd claude-code-agents-team-nation-of-elites
bash scripts/deploy_agents.sh
Done! All 74 agents are now in ~/.claude/agents/ and ready to use.
WSL2 note: The path ~/.claude appears in Windows Explorer as \\wsl.localhost\Ubuntu\home\<USER>\.claude.
git clone https://github.com/advisely/claude-code-agents-team-nation-of-elites.git
cd claude-code-agents-team-nation-of-elites
powershell -ExecutionPolicy Bypass -File scripts\deploy_agents.ps1
Done! All 74 agents are now in %USERPROFILE%\.claude\agents\ and ready to use.
Plugin Installation (Alternative)
Nation of Elites is a third-party plugin, so you add its marketplace first, then install from it (two steps):
# 1. Register this repo as a marketplace
/plugin marketplace add advisely/claude-code-agents-team-nation-of-elites
# 2. Install the plugin from it
/plugin install nation-of-elites@nation-of-elites
All 74 agents become available as scoped subagents (e.g. nation-of-elites:03_Engineering_Division:backend-developer), plus the 33 skills.
To hack on the plugin locally without installing, load it straight from a clone:
git clone https://github.com/advisely/claude-code-agents-team-nation-of-elites.git
claude --plugin-dir ./claude-code-agents-team-nation-of-elites
Manual Installation
git clone https://github.com/advisely/claude-code-agents-team-nation-of-elites.git
cd claude-code-agents-team-nation-of-elites
mkdir -p ~/.claude && cp -r agents ~/.claude/agents
git clone https://github.com/advisely/claude-code-agents-team-nation-of-elites.git
cd claude-code-agents-team-nation-of-elites
New-Item -ItemType Directory -Force -Path "$env:USERPROFILE\.claude\agents" | Out-Null
Copy-Item -Recurse -Force agents\* "$env:USERPROFILE\.claude\agents\"
Using the Agents
cd my-project
# Strategic planning - auto-spawns project-sponsor, program-manager, product-manager
claude "Initialize a new e-commerce project. Define goals, roadmap, and tech stack."
# Development - spawns chief-operations-orchestrator and specialists
claude "Build user authentication with React frontend and Django REST API backend"
# Code review - spawns code-reviewer with Semgrep SAST
claude "Review the payment integration code for security and best practices"
# Deployment - spawns devops-engineer and cloud-architect
claude "Setup production infrastructure with CI/CD pipeline on AWS"
Complete Agent Roster by Department
Executive Office
| Agent | Role | |
|---|---|---|
project-sponsor | Strategic vision, budget, ROI | |
program-manager | Portfolio oversight, cross-project |
Strategy & Planning
| Agent | Role | |
|---|---|---|
business-analyst | Requirements, gap analysis (opus) | |
functional-analyst | Process modeling, specs (opus) | |
product-manager | Vision, roadmap, priorities (opus) | |
solution-architect | Tech architecture, stacks (opus) | |
ux-ui-architect | UX research, interfaces |
Project Management
| Agent | Role | |
|---|---|---|
project-manager-scrum-master | Agile delivery, sprints |
Orchestration & Integration
| Agent | Role | |
|---|---|---|
chief-operations-orchestrator | Central coordinator | |
team-configurator | Stack detection, setup | |
integration-specialist | MCP, APIs, OAuth |
| Agent | Role | Model | |
|---|---|---|---|
api-architect | REST/GraphQL API design | opus | |
graphql-architect | Schemas, federation, subscriptions | opus | |
backend-developer | Server-side APIs, business logic | sonnet | |
frontend-developer | UI frameworks, responsive design | sonnet |
Python & Web Frameworks
| Agent | Skills Preloaded | Model | |
|---|---|---|---|
django-expert | django-patterns | sonnet | |
laravel-expert | laravel-patterns | sonnet | |
rails-expert | rails-patterns | sonnet |
JavaScript & Frontend Frameworks
| Agent | Skills Preloaded | Model | |
|---|---|---|---|
react-expert | react-patterns | sonnet | |
react-typescript-expert | react-patterns, typescript-patterns | sonnet | |
vue-expert | vue-patterns | sonnet | |
nextjs-expert | nextjs-patterns | sonnet |
Backend Languages
| Agent | Skills Preloaded | Model | |
|---|---|---|---|
java-expert | java-patterns | sonnet | |
go-expert | go-patterns | sonnet | |
rust-expert | -- | sonnet |
Data & Finance
| Agent | Skills Preloaded | Model | |
|---|---|---|---|
database-expert | -- | sonnet | |
financial-systems-expert | financial-trading-patterns | sonnet | |
crypto-api-developer | crypto-defi-patterns | sonnet |
UI & Styling
| Agent | Skills Preloaded | Model | |
|---|---|---|---|
tailwind-css-expert | tailwind-patterns | sonnet | |
antd-ui-developer | antd-patterns | sonnet |
CAD, BIM & Game Engines
| Agent | Skills Preloaded | Model | |
|---|---|---|---|
unreal-engine-expert | -- | sonnet | |
catia-design-expert | -- | sonnet | |
autodesk-cloud-construction-expert | -- | sonnet | |
revit-bim-expert | -- | sonnet |
| Agent | Role | Memory | |
|---|---|---|---|
code-reviewer | Security-aware PR reviews (opus) | project | |
code-archaeologist | Legacy codebase analysis (opus) | project | |
documentation-specialist | READMEs, API docs, guides (opus) | -- | |
performance-optimizer | Bottleneck diagnosis | -- | |
accessibility-specialist | WCAG 2.1 AA/AAA | -- | |
noe-general-purpose | Versatile multi-language | -- | |
noe-statusline-setup | Terminal/shell config | -- | |
output-style-setup | CLI output, ANSI colors | -- |
| Agent | Role | Skills | |
|---|---|---|---|
qa-test-planner | Test strategy, case design | -- | |
qa-engineer | Automated suites, CI/CD | pytest, semgrep, pipeline-quality | |
automated-test-scripter | Test framework implementation | pytest, semgrep | |
visual-regression-specialist | UI consistency, visual defects | -- |
| Agent | Role | Memory | |
|---|---|---|---|
cyber-sentinel | Vulnerability scanning, audits (opus) | project | |
storage-security-specialist | Data protection, encryption | -- |
| Agent | Role | Skills | |
|---|---|---|---|
devops-engineer | CI/CD, automation | github-actions, k8s, semgrep, pipelines | |
cloud-architect | Cloud infra, IaC | terraform-patterns | |
infrastructure-specialist | Servers, networking | -- | |
platform-engineer | Developer platforms, DevEx | -- | |
sre-specialist | SLO/SLI, incident mgmt | -- | |
observability-engineer | Metrics, logs, traces | -- | |
message-queue-specialist | Kafka, RabbitMQ, events | -- |
| Agent | Role | Model | |
|---|---|---|---|
ai-strategist | AI opportunities, vision | opus | |
data-scientist | EDA, modeling, ML prototyping | sonnet | |
data-engineer | ETL pipelines, data platforms | sonnet | |
ml-engineer | MLOps, model productionization | sonnet | |
aiops-specialist | ML model lifecycle, ops health | sonnet |
| Agent | Role | Model | |
|---|---|---|---|
mobile-architect | Cross-platform strategy | sonnet | |
ios-developer | Swift/SwiftUI, App Store | sonnet | |
android-developer | Kotlin/Compose, Play Store | sonnet |
| Agent | Role | Model | |
|---|---|---|---|
construction-ai-orchestrator | IMPARARIA AI-Stimate/Takeoff/Coordinate | sonnet |
| Agent | Role | Model | Skills | |
|---|---|---|---|---|
book-author | Drafting, ghostwriting, narrative voice | opus | humanizer | |
book-editor | Developmental editing, copy editing, proofread | opus | humanizer | |
publishing-specialist | KDP, IngramSpark, metadata, ISBNs | sonnet | humanizer | |
translation-localization-specialist | Multilingual translation with cultural fidelity | opus | humanizer |
| Agent | Role | Model | Skills | |
|---|---|---|---|---|
business-development-manager | Pipeline strategy, deal lifecycle, win/loss | opus | humanizer | |
lead-generation-specialist | Prospecting, outreach, lead scoring | opus | humanizer | |
proposal-architect | RFP response, proposals, compliance matrices | opus | humanizer | |
client-success-manager | Client lifecycle, retention, renewals | opus | humanizer | |
market-intelligence-analyst | Competitive intel, market analysis, pricing | opus | -- | |
social-media-strategist | Social content, paid ads, campaigns | opus | humanizer, adspirer |
Organization at a Glance
agents/
00_Executive_Wing/ (2) Strategic vision & portfolio management
01_Strategy_and_Planning/ (5) Architecture, requirements, UX design
02_Project_Management/ (1) Agile delivery & sprint management
03_Engineering_Division/ (31) Core dev + 19 framework specialists + 8 code excellence
04_Quality_Assurance/ (4) Testing strategy, automation, visual regression
05_SecOps_and_Infrastructure/ (9) Security, DevOps, cloud, SRE, observability
06_AI_and_ML_Division/ (5) Data science, ML engineering, AIOps
07_Orchestrators/ (3) Coordination, team config, MCP integration
08_Mobile_Development/ (3) iOS, Android, cross-platform architecture
09_Construction_Industry/ (1) Construction AI (IMPARARIA methodology)
10_Content_and_Localization/ (4) Books, publishing, translation, localization
11_Business_Development/ (6) BD pipeline, proposals, social media, market intel
----
74 agents total
Key Capabilities
Claude Agent SDK Features
| Feature | Description |
|---|---|
| Claude Opus 4.8 | opus alias resolves to claude-opus-4-8 — adaptive thinking, high default effort (all surfaces), sharper judgment (~4× fewer unremarked code flaws), better long-context & compaction recovery, high-res vision (2576px / 3.75MP) |
| Subagent Coordination | Parallel processing with isolated context windows (4.8 has better tool triggering; Dynamic Workflows can fan out to hundreds of verified subagents) |
| Dynamic Workflows (research preview) | Plan a task, then spin up hundreds of parallel verified subagents in a single Claude Code session |
| Agent Teams (experimental) | Multi-agent parallel work with shared task lists (CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1) |
| Mid-conversation system messages | role: "system" after a user turn (no beta header) — preserves prompt cache on long agentic loops |
| Task Budgets (beta) | Advisory token countdown across a full agentic loop (task-budgets-2026-03-13 beta header) |
| Context Compaction | Automatic summarization at 80% usage (47% token reduction) |
| Persistent Memory | 13 agents retain project knowledge across sessions via memory: project |
| Skills Preloading | 18 specialists preload matching skills at startup via skills: frontmatter |
| Permission Modes | 16 code-writing agents use permissionMode: acceptEdits for frictionless flow |
| MCP Integration | External services (Slack, GitHub, Jira, Figma, Sentry) via Model Context Protocol |
| Worktree Isolation | Subagents work in isolated git worktrees for parallel development |
| Official Plugins | Auto-detection and setup of Anthropic official plugins (GitHub, Slack, Jira, etc.) |
| Semgrep SAST | Automated security scanning via CLI and MCP plugin integration |
| Universal Pipelines | Stack-adaptive quality gates and full build pipelines |
Thinking Policies & Budgets
Scratchpad token budgets (internal reasoning) mapped to Opus 4.8 effort levels:
| Complexity | Tokens | Effort Level | Agents |
|---|---|---|---|
| High | 600-800 | xhigh / max | Solution, UX/UI, API, Cloud Architects |
| Medium | 400-600 | high | Business/Functional Analysts, QA Planner, AI Strategist |
| Medium-Low | 200-300 | medium | Framework Specialists, DevOps, Security |
| Low | 100-200 | low / medium | Backend/Frontend Devs, QA Engineer, Performance |
| Orchestration | <=300 | medium | Chief Operations Orchestrator, Team Configurator |
On Opus 4.8, effort defaults to high on all surfaces (API + Claude Code); use xhigh for hard design/architecture or large-codebase work, overriding per-agent via effort: frontmatter only when warranted.
Official Plugin Integrations
The deploy script auto-detects and offers to configure these official Anthropic plugins:
| Plugin | Category | Best For |
|---|---|---|
| GitHub | Development | code-reviewer, devops-engineer |
| Slack | Communication | chief-operations-orchestrator, program-manager, client-success-manager |
| Jira | Project Management | project-manager-scrum-master, product-manager |
| Linear | Project Management | product-manager, business-development-manager |
| Figma | Design | ux-ui-architect, frontend-developer |
| Sentry | Monitoring | sre-specialist, observability-engineer |
| Vercel | Deployment | nextjs-expert, devops-engineer |
| Firebase | Backend | mobile-architect, backend-developer |
| Supabase | Database | database-expert, backend-developer |
| Notion | Documentation | documentation-specialist, business-analyst, proposal-architect |
| Confluence | Documentation | functional-analyst |
| Asana | Project Management | program-manager |
Agent Skills System
33 custom skills + 9 official Anthropic skills with progressive 3-level loading.
How Skills Work
| Aspect | Agents | Skills |
|---|---|---|
| What | Team members (74 specialists) | Training manuals & toolkits |
| Where | ~/.claude/agents/ | ~/.claude/skills/ |
| Loading | Task-based spawning | Progressive disclosure |
Skills Library
| Skill | Purpose |
|---|---|
| PDF manipulation, forms, extraction | |
| docx | Word document creation/editing |
| pptx | PowerPoint presentations |
| xlsx | Excel with formulas and charts |
| mcp-builder | MCP server development |
| webapp-testing | Playwright UI testing |
| skill-creator | Interactive skill development |
| artifacts-builder | React/Tailwind components |
| canvas-design | Visual art creation |
| Skill | Framework | Preloaded By |
|---|---|---|
| django-patterns | Django ORM, DRF, auth | django-expert |
| react-patterns | React hooks, performance | react-expert, react-typescript-expert |
| vue-patterns | Vue 3 Composition API, Pinia | vue-expert |
| nextjs-patterns | Next.js 14 App Router, RSC | nextjs-expert |
| rails-patterns | Ruby on Rails, ActiveRecord | rails-expert |
| java-patterns | Spring Boot, JPA, microservices | java-expert |
| go-patterns | Concurrency, interfaces | go-expert |
| laravel-patterns | Eloquent, API resources | laravel-expert |
| Skill | Purpose | Preloaded By |
|---|---|---|
| security-audit | OWASP Top 10 checklist | cyber-sentinel |
| semgrep-sast | SAST scanning + MCP plugin | cyber-sentinel, code-reviewer, qa-engineer |
| pipeline-quality | Universal quality gate | code-reviewer, qa-engineer, devops-engineer |
| pipeline-full-build | Universal build pipeline | devops-engineer |
| github-actions | CI/CD templates | devops-engineer |
| kubernetes-deployment | K8s patterns, autoscaling | devops-engineer |
| terraform-patterns | IaC modules, state mgmt | cloud-architect |
| Skill | Category |
|---|---|
| humanizer | Writing Quality (removes AI writing patterns) |
| tailwind-patterns | UI/Styling |
| antd-patterns | UI/Styling |
| financial-trading-patterns | Specialized Domain |
| crypto-defi-patterns | Specialized Domain |
| pytest-patterns | Testing |
| silent-failure-audit | Quality Assurance |
| feature-workflow | Workflow Automation |
| quick-fix | Workflow Automation |
| pr-ready | Workflow Automation |
| Skill | Focus |
|---|---|
| nodejs-patterns | Async, streams, worker threads |
| python-patterns | Type hints, async, dataclasses |
| fastapi-patterns | DI, async DB, auth |
| express-patterns | Middleware, error handling |
| typescript-patterns | Advanced types, generics |
| vite-patterns | Build config, plugins |
| zustand-patterns | State management, slices |
For detailed skills documentation, see SKILLS.md.
Documentation
| Document | Purpose |
|---|---|
| README.md | Quick start and overview (this file) |
| CLAUDE.md | Project configuration and rule index |
| SKILLS.md | Agent Skills system documentation |
| CHANGELOG.md | Version history and release notes |
| MIGRATION_GUIDE.md | Upgrading from v1.x to v2.0+ |
| CONTRIBUTING.md | Community contribution guidelines |
| docs/SDK_COMPLIANCE_REPORT.md | 10/10 SDK compliance certification |
| docs/CONTEXT_ENGINEERING.md | Context management best practices |
| docs/rules/ | Modular rule files (7 files) |
Upgrading
git clone https://github.com/advisely/claude-code-agents-team-nation-of-elites.git
cd claude-code-agents-team-nation-of-elites
bash scripts/deploy_agents.sh # Linux/macOS/WSL2
# or
powershell -ExecutionPolicy Bypass -File scripts\deploy_agents.ps1 # Windows
All 74 agents support both natural language invocation (v2+) and explicit agent mentions (v1 legacy). No breaking changes.
Contributing
Contributions are welcome! See CONTRIBUTING.md for guidelines.
- Fork the project
- Create your feature branch (
git checkout -b feature/AmazingNewAgent) - Follow the Agent Development Guidelines
- Commit with conventional commits
- Open a Pull Request
Credits & References
| Resource | |
|---|---|
| Author | Yassine Boumiza -- GitHub |
| Repository | advisely/claude-code-agents-team-nation-of-elites |
| Anthropic Docs | Building Effective Agents |
| Agent Skills | Equipping Agents for the Real World |
| Agent SDK | Building Agents with Claude Agent SDK |
| Plugins | Claude Code Plugins Announcement |
| Agent Teams | Building a C Compiler with Agent Teams |
| Construction AI | IMPARARIA -- Progressive Deployment Methodology |
License
Distributed under the MIT License. See LICENSE.md for details.
If you find this project useful:
- Star the repository
- Share with your team
- Discuss / Report Issues
Nation of Elites v3.12.0 -- Claude Opus 4.8 | 74 Agents | 12 Divisions | 33 Skills | Dynamic Workflows | Task Budgets | Adaptive Thinking | BD & Content Wings | Official Plugins | Agent Teams | Semgrep SAST