Swarm Recommender

The Swarm Recommender (swarm_recommender.py, 13,520 bytes) analyses a project description and recommends an optimal swarm topology — how many agents, which archetypes, and what division structure to deploy.

This is the entry point for zero-to-running: describe what you want to build, and the recommender designs the team.

How It Works

Project Description
       │
       ▼
┌────────────────────────┐
│  Workload Analysis     │
│  - Complexity score    │
│  - Domain detection    │
│  - Skill requirements  │
└──────────┬─────────────┘
           │
           ▼
┌────────────────────────┐
│  Topology Selection    │
│  - Team size           │
│  - Division count      │
│  - Archetype mix       │
└──────────┬─────────────┘
           │
           ▼
┌────────────────────────┐
│  Blueprint Generation  │
│  - Hiring sequence     │
│  - Reporting chains    │
│  - Budget allocation   │
└────────────────────────┘

Workload Analysis

The recommender evaluates the project against several dimensions:

Topology Templates

Based on the analysis, one of several topology templates is selected:

Solo — Single agent for simple tasks

Owner → Builder

Duo — Builder + Auditor for quality-critical work

Owner → Builder → Auditor

Squad — Full engineering team

Owner → CEO → CTO
                ├── Architect
                ├── Builder (Frontend)
                ├── Builder (Backend)
                └── Auditor

Division — Large project with multiple departments

Owner → CEO
         ├── CTO → Architects + Builders
         ├── COO → Implementors
         └── CMO → Content Agents

Integration

The recommender integrates with:

• Blueprint Designer — Recommended topologies are converted into executable blueprints
• Budget Service — Each recommended topology includes a budget estimate
• Hire Approval — The recommended agents are queued for hiring upon approval
• Dashboard — Operators can review and modify the recommendation before deployment