nityeshaga/claude-home-base/explorable-explanations

Launch a separate subagent to act as a creative director. Have it review the code and visually inspect a few key pages, including the playables. Assume the structure, teaching, and writing are already excellent. Its job is to bring personality to the design.
 
Ask it to propose a few creative directions that feel specific to this topic. Think at two scales:
 
- The overall experience: like color, typography, backgrounds, texture, illustration style, motion, and recurring visual motifs that make everything feel part of the same world.
- Individual slides or groups of slides: distinctive treatments for key pages—an expressive character, a surprising SVG illustration, a playful background, a mascot, or a small visual joke that makes an idea stick.
 
Give it permission to be a radical design thinker with a tasteful restraint. What could make someone recognize this explorable from a single screenshot? Consider shaders or 3D effects combined with images where they would make the experience more interesting. It can use the openai-imagegen skill for image generation.
 
Keep the clarity and ease of use that already work. Let the personality come from the subject and the experience of learning it.
 
Have the creative director recommend its strongest ideas, with concrete examples of how it would transform a few existing pages. I want to be able to picture the result.

Launch a separate subagent whose job is to review the narrative flow of this whole explorable.
 
it needs to put itself in the shoes of a reader going through the material - say someone called Chad who is an impatient but curious teenager, he doesn't hate spending time on something but it needs to keep him engaged.
 
it needs to put itself in the shoes of a reader and point out all the slides when the flow gets disturbed - material goes too fast, or jumps to a new idea without setting the expectation or building up properly or curiosity gap wasn't established, maybe a slide jumps onto the next with no connection, or maybe it is too dense or maybe a concept didn't the time it deserved and it leaves Chad hungry to learn more about it or a promise made in an earlier slide goes unfulfilled.
 
so he's going to go through all the slides (just ask her to view the code) and give us pointed, blunt, specific feedback on any slide that doesn't work for him and why. he doesn't suggest solutions only specific problems and he doesn't mince his words.