Types give agents structural hints before opening documents

Type: note · Status: seedling

Agents are stateless — they start fresh every session with no memory of what they've read before. Context is finite and expensive. Without types, an agent must either load everything (wasteful) or guess what's relevant (error-prone).

Types solve this by giving agents structural hints before opening a document. A spec tells an agent it can implement from this. A structured-claim tells it there's a developed argument with evidence. An index tells it this is a navigation hub. These hints enable informed routing decisions: the agent reads a type and description, then decides whether to load the full document.

This is why description is the most important field — it's a retrieval filter, not a summary. The type tells the agent what kind of thing a document is; the description tells it which instance among documents of that kind is relevant to the current task. Together they let an agent narrow from hundreds of files to the few it needs without opening any of them.

The verifiability criterion is what makes this work: types must assert structural properties, not subject matter. "This is a design note" tells an agent nothing it can act on — every note in a design KB is about design. "This has Evidence and Reasoning sections" tells the agent it can extract a citable argument. The structural promise is what makes the routing decision informed rather than blind.

But these structural hints only work if the metadata exists reliably. That's the role of type-system enforcement — descriptions don't appear spontaneously; the type system is what makes them exist.

Relevant Notes:

Topics: