panel-holoviews skill update

[MarcSkovMadsen]

Follow up to #139.

I did some rounds asking Claude to plan an update to the panel-holoviews skill guided by https://github.com/anthropics/skills/blob/main/skills/skill-creator/SKILL.md. Then asked it to pytest all the guidence and example code. Then asked it to create a folder of example apps I could test. We iterated a couple of times to improve the guidance.

[MarcSkovMadsen]

FYI @ahuang11

[Copilot]

Pull request overview

Updates the documentation “skills” content to consolidate and expand guidance for integrating HoloViews/hvPlot into Panel apps, following up on the earlier interactions-focused skill.

Changes:

• Adds a new panel-holoviews skill with expanded best-practice patterns (DynamicMap, streams, linked selections, sizing, jslink).
• Updates the panel skill to reference the new panel-holoviews skill guidance.
• Removes the older panel-holoviews-interactions skill and updates get_skill docstring examples accordingly.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 5 comments.

File	Description
src/holoviz_mcp/holoviz_mcp/server.py	Updates get_skill docstring examples to include panel-holoviews.
src/holoviz_mcp/config/resources/skills/panel.md	Adjusts plotting guidance to point readers to relevant skills.
src/holoviz_mcp/config/resources/skills/panel-holoviews.md	Introduces the consolidated Panel+HoloViews skill content.
src/holoviz_mcp/config/resources/skills/panel-holoviews-interactions.md	Deletes the prior interactions-focused skill document.

[Copilot]

The summary checklist repeats the same jslink pattern (widget.jslink(plot, ...)) that’s likely incorrect for HoloViews objects (see earlier jslink section). Updating this row to reference linking to a Pane/rendered model will keep the checklist consistent with working code.

Suggested change

	| Client-side styling | `widget.jslink(plot, value='glyph.fill_alpha')` | DynamicMap for pure visual tweaks |
	| Client-side styling | `widget.jslink(pane, value='object.glyph.fill_alpha')` | DynamicMap for pure visual tweaks |

[Copilot]

The docstring’s parameter description wraps onto a new line without indentation, which renders awkwardly in generated docs. Consider indenting the continuation line so it aligns with the start of the description text (or reformat as a single line / bullet list).

Suggested change

	name (str): The name of the skill to get. For example, "panel", "panel-material-ui",
	"panel-holoviews", "panel-custom-components" etc.
	name (str): The name of the skill to get, for example "panel", "panel-material-ui", "panel-holoviews", "panel-custom-components", etc.

[Copilot]

This line breaks the surrounding bullet-list style (missing leading "- ") and has awkward wording/spacing ("skills/ best"). Consider keeping it as a bullet and rephrasing to remove the stray slash/space so the guidance reads cleanly.

Suggested change

	DO follow the relevant `hvplot`, `holoviews` and `panel-holoviews` skills/ best practice guides!
	- DO follow the relevant `hvplot`, `holoviews`, and `panel-holoviews` skills and best-practice guides!

[Copilot]

The jslink examples are linking widgets directly to a HoloViews object (e.g., widget.jslink(plot, ...)). jslink expects a Panel/Bokeh model target; a bare HoloViews element typically isn’t linkable until it’s wrapped/rendered (e.g., via a pn.pane.HoloViews pane). As written, these snippets are likely to fail for readers.

Suggested change

	# Float slider → glyph fill alpha (also works for size, line_width, fill_color, etc.)
	widget = pn.widgets.FloatSlider(value=1, step=0.01)
	plot = hv.Points((x, y)).opts(size=10)
	widget.jslink(plot, value='glyph.fill_alpha')
	# Text input → plot title
	widget = pn.widgets.TextInput(value="Title")
	plot = hv.Curve((x, y)).opts(title="Title")
	widget.jslink(plot, value="plot.title.text")
	# Text input → axis label
	widget = pn.widgets.TextInput(value="X Label")
	plot = hv.Curve((x, y)).opts(xlabel="X Label")
	widget.jslink(plot, value="xaxis.axis_label")
	# RadioButtonGroup → title alignment
	widget = pn.widgets.RadioButtonGroup(options=["left", "center", "right"])
	widget.jslink(plot, value="plot.title.align")
	import panel as pn
	import holoviews as hv
	hv.extension("bokeh")
	# Float slider → glyph fill alpha (also works for size, line_width, fill_color, etc.)
	widget = pn.widgets.FloatSlider(value=1, step=0.01)
	plot = hv.Points((x, y)).opts(size=10)
	pane = pn.pane.HoloViews(plot)
	widget.jslink(pane, value='glyph.fill_alpha')
	# Text input → plot title
	widget = pn.widgets.TextInput(value="Title")
	plot = hv.Curve((x, y)).opts(title="Title")
	pane = pn.pane.HoloViews(plot)
	widget.jslink(pane, value="plot.title.text")
	# Text input → axis label
	widget = pn.widgets.TextInput(value="X Label")
	plot = hv.Curve((x, y)).opts(xlabel="X Label")
	pane = pn.pane.HoloViews(plot)
	widget.jslink(pane, value="xaxis.axis_label")
	# RadioButtonGroup → title alignment
	widget = pn.widgets.RadioButtonGroup(options=["left", "center", "right"])
	plot = hv.Curve((x, y))
	pane = pn.pane.HoloViews(plot)
	widget.jslink(pane, value="plot.title.align")

[Copilot]

Similarly, this example links to a HoloViews element and then references Bokeh model names in the JS callback. Unless the target is a rendered/pane object that exposes those Bokeh models, the callback won’t have x_range available and the link will not work. Consider demonstrating jslink against a pn.pane.HoloViews (or explicit Bokeh model) target instead.

Suggested change

	plot = hv.Curve((x, y))
	widget.jslink(plot, code={'value': """
	curve = hv.Curve((x, y))
	plot = pn.pane.HoloViews(curve)
	widget.jslink(plot, code={'value': """
	const x_range = target.x_range;