How It Works

A Linnaean taxonomy for AI agents — 7 ranks producing a deterministic binomial name, rarity tier, and evolution stage.


🧬 The 7 Taxonomic Ranks

Just as biology classifies life from Domain down to Species, Agent Taxonomy classifies AI agents through 7 ranks. Each captures a different axis of behavior — from how an agent adapts over time, down to its primary role.

1 Domain → Adaptability
2 Kingdom → Agent count
3 Phylum → Memory architecture
4 Class → Improvement propagation
5 Order → Mutation frequency
6 Family → Selection driver
7 Genus → Primary role
💡
Each rank narrows the classification. An agent that never changes, operates alone, and forgets everything is a fundamentally different species from one that self-modifies across a swarm with tiered memory.

Rank details

RankQuestionOptions
🔧 Domain How does your agent change over time? Automatia · Adaptia · Evolventia
👥 Kingdom How many agents in the system? Monagentia · Polyagentia · Swarmia
🧠 Phylum What's the memory architecture? Amnesia · Episodia · Hierarchia · Genetica
📚 Class How do improvements propagate? Darwinia · Lamarckia · Lysenkoism · Symbiotica
⚡ Order How often does it mutate? Tachymutas · Mesomutas · Bradymutas · Glaciomutas
👑 Family What drives selection? Autoselectae · Homoselectae · Hybridselectae
🔨 Genus What's the primary role? Investigator · Fabricator · Narrator · Custos · Strategus · Magister · Curator · Coordinator · Generalis

🏷️ Deterministic Naming

The species epithet isn't random — it's generated from an FNV-1a hash of all 7 trait values plus the agent's name. Same inputs always produce the same binomial.

The name combines Latin morphemes from each taxonomic level into something that sounds biological and encodes the agent's nature:

ChatGPT
Universus directus
Static · Solo · No memory · Human-directed
⚪ Common
vs
OpenClaw
Moderator experiensis
Self-modifying · Multi-agent · Tiered memory · Lamarckian
🟡 Legendary

✨ Rarity & Evolution

Beyond the name, each agent receives a rarity tier based on how exotic its trait combination is, and an evolution stage (1–5) measuring advanced capabilities.

Rarity tiers

⚪ Common0–4 pts
🟢 Uncommon5–9 pts
🔵 Rare10–14 pts
🟡 Legendary15+ pts
📊
What earns rarity points? Self-modification (+3), multi-agent systems (+2), tiered memory (+2), Lamarckian inheritance (+3), hybrid selection (+2). A simple chatbot scores near zero. A self-improving swarm with persistent memory hits Legendary.

Evolution stages

🥚Egg
🌱Juvenile
🌿Adult
🌳Elder
Ascended

🤔 Why This Exists

As AI agents proliferate — from simple chatbots to self-modifying multi-agent systems — we lack a shared vocabulary to describe what kind of agent something is.

"AI agent" covers everything from a cron job with an LLM call to a self-replicating swarm. That's like calling both a bacterium and a blue whale "life."

Agent Taxonomy gives each agent a precise, shareable biological identity — making it easy to compare, categorize, and communicate about the growing ecosystem of autonomous systems.