Browse / Software development / ai-agent-workforce
ai-agent-workforce
muhamadto
Deploy AI agent teams as code.
Open muhamadto/ai-agent-workforce → Does not match the four rules.
Does not say
no agent config and no instruction file at the root.
In its own words
Ansible automation for multi-model AI agents with extensible integrations, task orchestration, and workspace setup
Deploy Claude Code agent teams as code. Ansible automation for AI agent configurations, skills, and workspace management.
Overview
AI Agent Workforce is an Ansible-based automation tool for deploying and managing Claude Code agent teams. It provides infrastructure-as-code for agent configurations, skills, and integrations.
Key Features
- Specialized Agents - Pre-configured expert agents (backend, frontend, security, QA, BA, etc.)
- Shared Skills - Reusable slash-command skills deployed via symlink
- Extensible Integrations - MCP (Model Context Protocol) and custom plugins
- Idempotent Operations - Install and uninstall agents cleanly
- Security-First - Safe configuration management with environment-based secrets
Quick Start
Prerequisites
- macOS (Darwin)
- Ansible 2.9+
- Python 3.8+
Installation
- Deploy all agents locally:
- Deploy specific components using tags:
Uninstallation
Remove all agents and configurations:
Architecture
Project Structure
The claude role deploys Claude Code agents to ~/.claude/ (default). The opencode role deploys opencode configuration to ~/.config/opencode/ and is opt-in via --tags opencode.
Agents are slim personas (standards, workflow, posture); their domain knowledge lives in lazy-loaded knowledge skills (e.g. java-spring-engineering, auth-engineering) that agents pull in on demand via their skills: frontmatter.
The repository's own description, then 38% of its README. Read the rest on GitHub ↗
Read from the repository's git tree on 2026-08-17.