Agent Team Topologies
This directory contains 8 topology pattern cards – reusable team configurations for Claude Code agent teams. Each pattern describes a reusable approach to structuring multi-agent collaboration for a specific class of problem.
Pick the topology that matches your work shape, then use the spawn prompt inside to get started immediately.
Note: Quality-Gated is a composable topology – it layers on top of any other pattern to enforce completion standards via hooks. You can combine it with any topology below.
All topologies are primitives. Quality-Gated is the most explicit overlay, but any teammate slot can itself become a topology. A reviewer in Feature Pod can spawn a Review Board. An explorer can fan out sub-explorers. See Composing Topologies for recipes.
Topology Comparison
| Topology | Best For | Team Size | Cost | Parallelism |
|---|---|---|---|---|
| Parallel Explorers | Discovery, research, codebase mapping | 3-5 | Low | High |
| Review Board | Code review with distinct lenses | 4 | Low | High |
| Competing Hypotheses | Ambiguous bugs, architectural decisions | 4-7 | Medium | Medium |
| Feature Pod | Cross-layer feature delivery | 4 | Medium | High |
| Risky Refactor | High-risk changes needing plan approval | 4 | Low | Low (sequential) |
| Orchestrator-Only | Pure coordination, lead never codes | 3-6 | High | High |
| Quality-Gated | Enforcing completion standards | Any | Overlay | Varies |
| Task Queue | Many small independent tasks | 4-9 | High | Very High |
Cost legend: Low
Medium
High
Overlay