Example: Hello Agent

Document ID: EX-001
File Path: docs/16-examples/hello-agent.md
Version: 1.0.0
Status: Draft
Owner: Developer Relations Team
Last Updated: 2026-06-27

1. Goal

Define and run the simplest possible agent — no tools, no memory — to learn the core loop: define → run → observe.

2. Define

agents/hello.yaml:

apiVersion: agent.wovyr.io/v1
kind: Agent
metadata:
  name: hello
spec:
  model_selector: { capability: chat, class: fast }
  instructions: |
    You are a friendly assistant. Greet the user and answer briefly.

The model selector lets the LLM Gateway pick a fast model; no model is pinned.

3. Run (Local)

wovyr agents run --local -f agents/hello.yaml \
  --input '{"message":"Hi, who are you?"}' --stream

--stream shows the run as it happens (run lifecycle):

start  · model: <fast model> (provider: ...)
delta  · "Hi! I'm a friendly assistant ..."
done   · tokens: 48, cost_usd: 0.0001

4. Run (Remote)

Publish, then run against a server:

wovyr agents create -f agents/hello.yaml
ID=$(wovyr agents list -o json | jq -r '.data[]|select(.name=="hello").id')
wovyr agents publish "$ID"
wovyr agents run "$ID" --input '{"message":"Hello!"}'

5. Observe

Usage: every response includes tokens + cost (token management).
Trace: the run produces a trace (gateway → provider) viewable in Agent Studio.

6. Next Steps

Add retrieval → RAG Agent
Add tools → Code Agent
Orchestrate multiple steps → Customer Support

8. Revision History

Version	Date	Description
1.0.0	2026-06-27	Initial Hello Agent example