Wxrkr

Agentic AI Task Execution Platform for Business

Wxrkr is an intelligent task automation platform that orchestrates AI agents to execute business workflows. It connects to multiple task sources, leverages business context through MCP and RAG servers, and uses Mermaid diagrams to define strict, documented agent workflows.

Overview

Wxrkr transforms business processes into executable AI agent workflows. The platform:

• Ingests tasks from multiple sources - GitHub issues, webhooks, APIs, manual input, and more
• Provides business context - Connects to MCP (Model Context Protocol) servers and RAG (Retrieval-Augmented Generation) systems for domain-specific knowledge
• Defines workflows visually - Uses Mermaid diagrams to create strict, documented agent execution paths
• Executes autonomously - AI agents process tasks in isolated LXD containers with full business context
• Tracks everything - Real-time progress monitoring, execution logs, and outcome tracking

Key Concepts

Agentic AI Workflows

Unlike simple automation, Wxrkr's AI agents make intelligent decisions within defined guardrails. Mermaid diagrams specify:

• Decision points and conditional logic
• Required context and data sources
• Validation and approval gates
• Error handling and rollback procedures

Multi-Source Task Ingestion

Tasks flow into Wxrkr from:

• GitHub Issues - Automatic processing of labeled issues
• Webhooks - External system integrations
• API Endpoints - Programmatic task submission
• Manual Creation - User-initiated tasks through the dashboard

Business Context Integration

Agents access relevant business knowledge through:

• MCP Servers - Standardized protocol for exposing business data, APIs, and tools to AI agents
• RAG Systems - Vector databases with company documentation, policies, and historical decisions
• Project Repositories - Direct access to codebases and version control

Isolated Execution

Each task runs in a dedicated LXD container with:

• Clean, reproducible environment
• Network access for API calls and integrations
• Secure credential management
• Complete execution logging

Features

• Visual Workflow Designer - Create agent workflows using Mermaid diagram syntax
• Multi-Agent Orchestration - Deploy specialized agents for different task types
• Dynamic Scaling - Configurable worker pools that scale with demand
• Real-time Monitoring - Live progress updates via Livewire
• Context-Aware Execution - Agents query MCP/RAG servers for business-specific knowledge
• GitHub Integration - OAuth, repository access, PR creation, and issue management
• Audit Trail - Complete logging of agent decisions and actions

Tech Stack

• Laravel 12 - Backend framework
• Livewire 3 & Flux UI - Reactive frontend components
• SQLite - Database (easily swappable)
• LXD - Container runtime for isolated execution
• Anthropic Claude - Primary AI model for agent intelligence
• Queue Workers - Background task processing
• Mermaid - Workflow diagram specification

Installation

# Clone the repository
git clone https://github.com/FoleyBridge-Solutions/wxrkr.git
cd wxrkr

# Install dependencies
composer install
npm install

# Setup environment
cp .env.example .env
php artisan key:generate

# Create database
touch database/database.sqlite

# Run migrations
php artisan migrate

# Build assets
npm run build

# Start the application
php artisan serve

Configuration

Key environment variables:

APP_NAME=Wxrkr
APP_URL=your-app-url

# Database
DB_CONNECTION=sqlite

# GitHub OAuth
GITHUB_CLIENT_ID=your-client-id
GITHUB_CLIENT_SECRET=your-client-secret
GITHUB_REDIRECT_URI=your-callback-url

# Anthropic Claude
ANTHROPIC_API_KEY=your-api-key

# LXD Configuration
LXD_SOCKET=/var/snap/lxd/common/lxd/unix.socket

Queue Workers

Configure worker count for concurrent task execution:

# Start queue workers
php artisan queue:work --sleep=3 --tries=3 --timeout=600

# Or use supervisor for production (recommended)
# Config: /etc/supervisor/conf.d/wxrkr-worker.conf

Creating Agent Workflows

Define workflows using Mermaid diagrams in your project settings:

graph TD
    A[Receive Task] --> B{Check Context}
    B -->|MCP Available| C[Query Business Data]
    B -->|RAG Available| D[Search Documentation]
    C --> E[Execute with Context]
    D --> E
    E --> F{Validation}
    F -->|Pass| G[Create PR]
    F -->|Fail| H[Request Review]
    G --> I[Complete]
    H --> I

Loading

The platform parses these diagrams to guide agent execution and enforce workflow compliance.

Use Cases

• Automated Code Reviews - Agents review PRs against company standards from RAG
• Issue Resolution - Agents fix bugs by querying MCP for system state
• Documentation Generation - Agents create docs based on code and business context
• Data Processing - Agents transform data following workflow diagrams
• Compliance Checks - Agents validate work against regulatory requirements

Architecture

Task Sources → Wxrkr Queue → Agent Scheduler
                              ↓
                    ┌─────────┴──────────┐
                    ↓                    ↓
              MCP Servers          RAG Systems
                    ↓                    ↓
              LXD Container (Agent Execution)
                    ↓
              Results & Logging

License

Proprietary - FoleyBridge Solutions

Support

For support, contact: dev@foleybridge.solutions

Roadmap

• Visual Mermaid diagram editor
• Additional AI model providers
• Webhook management UI
• Built-in RAG server
• Advanced workflow templates
• Multi-tenant support