What it is
Xiao-K Skill is an installable AI agent skill.
It uses one original character, Xiao-K, to turn article ideas, process steps and conflicts into publishable hand-drawn covers, illustrations, state diagrams and short comics.
Xiao-K is never just decoration.
In every frame, the character must perform a core action: push, carry, repair, test, guard or share.
Who is Xiao-K
Xiao-K is a geometric workhorse always short on time:
- yellow headphones
- sleepy horizontal eyes
- clock tail
- spring legs
Not a tech genius — an ordinary actor who gets stuck, reworks, and still finishes the job.
That is why the character fits stories about normal people getting real work done with AI.
Problems it fixes
Most AI illustration fails are not about beauty:
- the character drifts every time
- the image does not carry the idea
- one image is stretched across every platform
- the character stands still and does nothing
The skill uses identity anchors, scene grammar and platform layout rules to make the process repeatable.
Workflow
- Identify channel, topic, count and text needs
- Extract the idea / action / conflict / result worth visualizing
- Choose structure, then one core action and at most two props
- Write a short shot config
- Generate one image at a time and QA for consistency
Where it fits in the content workflow
writing / practice notes
→ Xiao-K skill visual breakdown
→ multi-platform asset pack
→ publish and archiveIt is the visual execution layer of the content pipeline, not a full writing agent.