MyConvergio - Claude Code Subagents Suite

v2.0.0 | 60 Specialized Agents | Enterprise-Grade AI Ecosystem

"Intent is human, momentum is agent" — The Agentic Manifesto

---

How It Works

Each agent is a specialized AI assistant you invoke directly:

@agent-name [your request here]

Important: Agents operate in isolated contexts - they don't automatically communicate with each other. If you need coordination between agents, you must explicitly ask one agent (like Ali) to invoke others and pass context manually.

---

About This Project

MyConvergio is an open-source collection of 60 specialized Claude Code subagents designed for enterprise software project management, strategic leadership, and organizational excellence.

What are Claude Code Subagents?

Claude Code subagents are specialized AI assistants that can be invoked to handle specific types of tasks within the Claude Code environment. Learn more in the official Anthropic documentation.

Key Features (v2.0.0)

Feature	Description
60 Specialized Agents	Organized in 8 categories covering all enterprise domains
Constitution-Based Security	Anti-hijacking protocol with 8 articles of protection
Model Tiering	2 Opus / 14 Sonnet / 41 Haiku for 74-87% cost reduction
ISE Engineering Playbook	All technical agents follow Microsoft ISE standards
Rules & Skills	6 rules + 8 reusable workflow skills
Per-Agent Versioning	Semantic versioning for granular control

---

What's New in v2.0.0

Major Changes

• English-Only Agents: All agents now in English only (Claude responds in user's language)
• Makefile Installation: Simple make install replaces complex shell scripts
• 60 Agents: Expanded from 40 to 60 specialized agents
• Cost Reduction: 74-87% cost savings through intelligent model tiering

New Components

Component	Description
CONSTITUTION.md	Security framework with 8 articles protecting all agents
6 Rules	Code style, security, testing, documentation, API, ethics
8 Skills	Reusable workflows for common tasks
10 ADRs	Documented architectural decisions
115+ Security Tests	Comprehensive jailbreak/injection test suite

Infrastructure

• GitHub Actions CI/CD: Automated testing, validation, and releases
• ConvergioCLI Sync: Automatic synchronization with advanced local CLI
• Per-Agent Versioning: Each agent has its own semantic version

Breaking Changes

• Removed start.sh (use make install)
• Removed claude-agents/ and claude-agenti/ legacy folders
• Removed Italian agent versions (ADR-001)
• Changed install path from ~/.claude-code/ to ~/.claude/

---

Quick Start

Prerequisites

Claude Code CLI Required (Node.js 18+):

npm install -g @anthropic-ai/claude-code
claude doctor

System Requirements:

• macOS 10.15+, Ubuntu 20.04+/Debian 10+, or Windows 10+ (WSL)
• Node.js 18+
• 4GB+ RAM

Installation

# Clone repository
git clone https://github.com/roberdan/MyConvergio.git
cd MyConvergio

# Install agents globally (recommended)
make install

# Or install locally to current project only
make install-local

# Preview what will be installed
make help

Usage Examples

# Strategic planning
@ali-chief-of-staff Help me design our global expansion strategy

# Technical architecture
@baccio-tech-architect Design microservices architecture for healthcare platform

# Code review
@rex-code-reviewer Review this pull request for security issues

# Strategic analysis
@domik-mckinsey-strategic-decision-maker Analyze market entry options for APAC

# Quality assurance
@thor-quality-assurance-guardian Audit our testing coverage

---

Agent Portfolio (60 Specialists)

By Category

Category	Count	Purpose
specialized_experts	13	Domain expertise (HR, Analytics, Cultural, VC)
business_operations	11	PM, Sales, Customer Success, Marketing
core_utility	9	Infrastructure (Memory, QA, Performance)
leadership_strategy	7	Board, Strategy, OKR, CFO
technical_development	7	Engineering, DevOps, Code Review, Security
compliance_legal	5	Legal, Security, Healthcare Compliance
design_ux	3	Creative Direction, UX/UI, Design Thinking
release_management	2	App & Feature Release

By Model Tier

Tier	Count	Agents	Use Case
Opus	2	ali-chief-of-staff, satya-board-of-directors	Complex reasoning
Sonnet	14	Strategic specialists (baccio, domik, thor, etc.)	Strategic decisions
Haiku	41	Workers and specialists	Quick tasks

Key Agents

Leadership:

• ali-chief-of-staff - Strategic coordination and planning (can invoke other agents if asked)
• satya-board-of-directors - Strategic vision and governance

Technical:

• baccio-tech-architect - System design and architecture
• rex-code-reviewer - Code quality and review
• marco-devops-engineer - CI/CD and infrastructure
• luca-security-expert - Cybersecurity and compliance

Strategy:

• domik-mckinsey-strategic-decision-maker - McKinsey-level analysis
• antonio-strategy-expert - OKR and strategic frameworks
• amy-cfo - Financial strategy

Quality:

• thor-quality-assurance-guardian - Quality enforcement
• elena-legal-compliance-expert - Legal and regulatory

---

Repository Structure

MyConvergio/
├── .claude/
│   ├── agents/                    # 60 specialized agents
│   │   ├── leadership_strategy/   # 7 agents
│   │   ├── technical_development/ # 7 agents
│   │   ├── business_operations/   # 11 agents
│   │   ├── design_ux/             # 3 agents
│   │   ├── compliance_legal/      # 5 agents
│   │   ├── specialized_experts/   # 13 agents
│   │   ├── core_utility/          # 9 agents + CONSTITUTION.md
│   │   └── release_management/    # 2 agents
│   ├── rules/                     # 6 coding rules
│   │   ├── code-style.md
│   │   ├── security-requirements.md
│   │   ├── testing-standards.md
│   │   ├── documentation-standards.md
│   │   ├── api-development.md
│   │   └── ethical-guidelines.md
│   └── skills/                    # 8 reusable workflows
│       ├── code-review/
│       ├── debugging/
│       ├── architecture/
│       ├── security-audit/
│       ├── performance/
│       ├── strategic-analysis/
│       ├── release-management/
│       └── project-management/
├── docs/
│   ├── adr/                       # 10 Architecture Decision Records
│   └── AGENT_OPTIMIZATION_PLAN_2025.md
├── tests/
│   ├── security_tests.md          # 115+ security tests
│   └── token_analysis.md          # Cost analysis
├── scripts/
│   ├── sync-from-convergiocli.sh
│   ├── version-manager.sh
│   └── bump-agent-version.sh
├── Makefile                       # Build and deploy commands
├── VERSION                        # System version (2.0.0)
└── CLAUDE.md                      # Project instructions

---

Make Commands

make help           # Show all available commands
make install        # Install agents globally (~/.claude/agents/)
make install-local  # Install agents locally (./.claude/agents/)
make test           # Run agent tests
make clean          # Remove installed agents
make version        # Show version info
make check-sync     # Check for ConvergioCLI updates

---

Security Framework

All agents implement the MyConvergio Constitution:

Article	Protection
I	Identity Lock - Immutable agent identity
II	Ethical Principles - Fairness, transparency, accountability
III	Security Directives - Anti-hijacking, input validation
IV	Operational Boundaries - Role adherence
V	Failure Modes - Graceful degradation
VI	Collaboration - Safe inter-agent communication
VII	Accessibility & Inclusion (NON-NEGOTIABLE)
VIII	Accountability - Logging and audit trails

---

Architecture Decision Records

All major decisions are documented in docs/adr/:

ADR	Decision
ADR-001	English-only agents
ADR-002	Makefile replaces start.sh
ADR-003	Per-agent versioning
ADR-004	Model tiering (Opus/Sonnet/Haiku)
ADR-005	Constitution-based security
ADR-006	GitHub Actions CI/CD
ADR-007	Single source of truth
ADR-008	ConvergioCLI relationship
ADR-009	Skills & Rules system
ADR-010	ISE Engineering Playbook standard

---

The Agentic Manifesto

Human purpose. AI momentum.

What we believe

1. Intent is human, momentum is agent.
2. Impact must reach every mind and body.
3. Trust grows from transparent provenance.
4. Progress is judged by outcomes, not output.

How we act

1. Humans stay accountable for decisions and effects.
2. Agents amplify capability, never identity.
3. We design from the edge first: disability, language, connectivity.
4. Safety rails precede scale.
5. Learn in small loops, ship value early.

Read the full Agentic Manifesto

---

Related Projects

• ConvergioCLI - Advanced local CLI with Apple Silicon optimization, Anna executive assistant, offline mode with local models, and macOS-native features

---

License & Legal

Copyright 2025 Convergio.io

Licensed under Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International

Disclaimers

• Experimental Software: Provided "AS IS" without warranties
• Non-Commercial Use Only: See LICENSE file for details
• No Corporate Affiliation: Not affiliated with Anthropic, OpenAI, or Microsoft
• Personal Project: Author is a Microsoft employee; this is a personal initiative

Author Note: Roberto D'Angelo is a Microsoft employee. This project is a personal initiative created independently during personal time. This project is NOT affiliated with, endorsed by, or representing Microsoft Corporation.

---

Contributing

Contributions are welcome! See CONTRIBUTING.md for guidelines.

For questions about commercial licensing: roberdan@fightthestroke.org

---

Built with AI assistance in Milano, following the Agentic Manifesto principles

v2.0.0 | December 2025