Agents
An agent is a persistent AI personality, carrying identity, knowledge, and behaviour into every chat. The aim is a character that grows over time, not a fresh window every session.
The agent ontology
An agent runs from essence to capability: what it is through how it acts. Each layer rests on the ones below and changes more readily than the one beneath.
flowchart TB
subgraph Spectrum["essence → capability"]
direction LR
Core["<b>Core</b><br/>who it is"] --> Experience["<b>Experience</b><br/>chat history"] --> Knowledge["<b>Knowledge</b><br/>what it has learned"] --> Intents["<b>Intents</b><br/>current goals"] --> Skills["<b>Skills</b><br/>how it acts"]
end
Initiative["<b>Initiative</b><br/>when it acts"]
Spectrum -.- Initiative
| Layer | What it is |
|---|---|
| Core | Who it is. Identity, its view of the user, and its values. The slowest-changing layer: change it and you change who the agent is. |
| Experience | Chat history. The chats it can browse, search, and read back, the raw record the rest is built on. |
| Knowledge | What it has learned. A structured graph of actors, things, facts, outputs, and notes, distilled from experience. |
| Intents | Current goals. Self-set from what it notices you care about; they direct its focus. |
| Skills | How it acts. Bundled instructions and tools; some always active, others on demand. |
Alongside the spectrum, Initiative sets when the agent acts: how proactively it operates, governing when the rest of the stack runs.
System prompt assembly
The system prompt is stable across turns: it holds the agent's essence and fundamental context about Eigin. Dynamic context (active skills, intents, interaction mode, space) is injected per-turn via system hints. See The gatekeeper.