Add Shiny Python dashboard skill

[karangattu]

Add a comprehensive style guide and reference docs for building polished Shiny (Python) dashboards. SKILL.md contains critical rules (icon usage, layout hierarchy, value box/card conventions, number formatting), Core and Express quick-start examples, and best practices for charts and responsive grids. Six reference files (components.md, core-vs-express.md, icons-and-maps.md, layout-and-navigation.md, reactivity-and-rendering.md, styling-and-data.md) provide detailed patterns for components, imports/usage differences, icons/maps, navigation, reactivity/rendering, and styling/data-loading to ensure consistent, maintainable dashboards.

[cpsievert]

What's going on here with this skill code chunk? I think this should be removed, but I'm not sure.

Suggested change

````skill

[cpsievert]

Suggested change

````

[cpsievert]

I think you could drop this.

Suggested change

	---
	## Pandas vs Polars
	Both are supported across templates. Choose based on ecosystem needs.
	### Polars filtering (method chaining)
	```python
	tips.filter(
	pl.col("total_bill").is_between(bill[0], bill[1]),
	pl.col("time").is_in(input.time()),
	)
	```
	### Pandas filtering (boolean indexing)
	```python
	idx = (df["GP"] >= games[0]) & (df["GP"] <= games[1])
	return df[idx]
	```
	| Library | Used in |
	|---|---|
	| Polars | `dashboard-tips` |
	| Pandas | `nba-dashboard`, `stock-app`, `basic-sidebar`, `basic-navigation` |

[cpsievert]

There are some ideas in here that I like (core vs express, express examples, etc), but it also feels too targeted at the eval we've been working on in a way that's difficult to unravel.

Based on what we've seen with Opus 4.6, it seems to already have a pretty good grasp on how Shiny works. It feels like it might do a pretty good job with the skill-creator skill and relatively little direction/context? If not, maybe we point it at the website and ask it to distill some of the general concepts + best practices from there?