Task 9 - Commands & Expected Output

Commands and expected output style

Use plain markdown code fences — no shortcode nesting. Follow this style for every command/output pair:

  • No $ prompt in the command.
  • One command per block, preceded by a sentence that says what it does.
  • Introduce the output with “The output is similar to:”.
  • Use [...] for elided lines in long output.
  • Reserve tabs for genuine alternatives (e.g. bash vs PowerShell) — not for hiding output.

run= targets

Add {run="<target>"} to a bash, sh, shell or text fence to get a coloured header badge. Five targets get a fixed colour and label — bastion, local, pod, browser (each reads “Run on: <Target>”) and chatbot (reads “Ask the chatbot”, see below); anything else gets a neutral badge.

Check the pods on the cluster:

Run on: Bastion
kubectl get pods

The output is similar to:

Expected output
NAME                     READY   STATUS    RESTARTS   AGE
ollama-7d4f9c6b78-abcde  1/1     Running   0          4m
agent-6b8d5f9c7d-fghij   1/1     Running   0          4m

Run a command on your local machine:

Run on: Local
docker ps

Exec into a pod:

Run on: Pod
kubectl exec -it ollama-7d4f9c6b78-abcde -- /bin/sh

Open the app in a browser:

Run on: Browser
open http://localhost:8080

An unrecognised target still renders, with a neutral badge:

Run on: Jumphost
whoami

chatbot blocks — for chat-UI prompts

run= also works on text fences, for prompts typed into a workshop’s chat UI rather than run in a shell. chatbot is a fifth standard target: header reads “Ask the chatbot” (not “Run on: …”), orange badge, FortiAI-Assist icon.

Enter this in the chat box:

Ask the chatbot
Who is in the Engineering department?

Custom label, icon and color

Every block’s header text, icon and colour are author-editable, on any target, with the label, icon and color attributes — for a one-off badge that isn’t one of the five standard targets:

Ask FortiAI-Assist
Summarize this alert and recommend a remediation.
  • label — replaces the header text entirely.
  • icon — terminal (default for most targets), fortiai (the mark chatbot uses by default), or any Font Awesome class string, e.g. fas fa-robot.
  • color — any CSS colour (hex, named, rgb()) for the header background.

output fences

Plain output, no syntax colouring, no copy button:

Expected output
qwen2.5:3b

Optional lang="json" highlights inside the panel:

Expected output
{
  "models": ["qwen2.5:3b"]
}

Optional collapse="true" wraps long output in the theme’s expand widget (the value must be quoted — Hugo’s fence-attribute parser does not accept a bare collapse=true):

Run on: Bastion
kubectl describe pod ollama-7d4f9c6b78-abcde

The output is similar to:

Expected output
Expected output
Name:         ollama-7d4f9c6b78-abcde
Namespace:    ai101
[...]
Status:       Running

title= composes with run=

For a multi-command step, keep Relearn’s title alongside the run-on badge:

Run on: Bastion
​
helm upgrade --install ai101 ./helm/ai101 -f values-lab1.yaml

Fallback (no CentralRepo image needed)

If the run=/output render hooks aren’t in the image you’re building against yet, this renders a titled block today with zero CentralRepo changes:

​
kubectl get pods
​
NAME   READY   STATUS