[idea:class-reference]reference2026-08-20

HTML class reference

You can also style more granularly with your own CSS using the classes below. A trailing * below stands for a family whose members are named in the description.

Class	What it is
`.idea-box`	An idea’s own block where it was written, carrying the left rule and the padding every rookery block nested inside it measures from.
`.idea-head`, `.idea-tab`	The heading group, and the short top rule the name straddles at the card’s corner.
`.idea`, `.idea-title`	The idea’s heading element—which carries the anchor an `@idea:etal` fragment resolves to—and the title text inside it.
`.idea-label`, `.idea-date`	The `[idea:etal]` permalink and the date beside it, in a heading, a window summary or prose, wherever either is shown.
`.idea-tag`, `.idea-tag-<tag>`	A tag pill in the hat, and one extra class per tag the idea carries. The second reaches every surface that names the tag—the idea’s heading, its box, its outline row, its pill, and a search result’s chip, which JavaScript builds in the browser—so a single rule of your own styles a tag everywhere it shows up.
`.idea-ref`	An `@idea:other` reference in prose. See referencing ideas.
`.idea-window`	A transclusion, wearing the same rule and indent as `.idea-box`; the second class `.idea-window-plain` opts a bare `#idea-body` out of that box.
`.idea-window-*`	The fold: `-details` the `<details>`, `-summary` the row you click, `-title` the name in that row, `-body` what folds away under it. The date beside the title is `.idea-date`, the same class it wears in a heading.
`.idea-window-more*`	The second fold a `limit:` makes inside the body: `.idea-window-more` its `<details>`, `-more-summary` the shown blocks you click to see the rest, and `.idea-window-ellipsis` the `…` at the end of them.
`.idea-outline*`	A page’s outline: `.idea-outline` the list at each level of nesting, `-title` its “Contents” label, `-row` one row per idea.
`.idea-footnote*`	An idea’s own footnotes, carried on every surface it appears on: `.idea-footnotes` the block, `-title` its label, `.idea-footnote-list` and `.idea-footnote` its entries.
`.idea-fn-*`	`-ref` the superscript mark in prose, `-backlink` the way back up to it from the entry.
`.idea-references`, `.idea-page-refs`	The bibliography an idea’s citations draw from, and the page’s own for citations written outside any idea.
`.idea-footer*`	On an idea page: `.idea-footer` the ruled-off apparatus under the note, `-title` the label on each of its sections.
`.idea-context`, `.idea-backlinks`	That footer’s two sections: where the note was written, and the notes pointing at it.
`.idea-page-list`, `.idea-page-row`	A list of pages in either section, and one row in it—a page cannot fold, so it wears a window’s shape without being one.
`.idea-row*`	One row of an `#idea-row` list: `.idea-row` the `<li>`, `-when` its date cell, `-title` the title, `-cell` each extra cell, `-badges` the chip strip at the end. `@rookery/search` and `@rookery/timeline` draw the same row, so a rule of yours here reaches all three.
`.idea-index-count`	The “n ideas” line on the minted index page.
`.idea-dup-warning`	The build-time notice shown when two ideas claim the same name.

Every one of those elements also carries a data-rookery attribute naming the same thing without the prefix—.idea-box is [data-rookery="box"], .idea-window-summary is [data-rookery="window-summary"]—and it is the attribute, not the class, that rookery’s own stylesheet selects on. The two say the same thing, so write whichever reads better in your CSS; the attribute is there because css-prefix moves the classes and nothing should have to follow them to find an element. .idea-window-plain is the one that does not repeat itself: a plain window is [data-rookery="window"] wearing data-rookery-plain besides.

Anything that names a tag—a card, a window, a pill, an outline row, an index row, a badge—carries one more, data-rookery-tags, listing those tag names, minus any made invisible by invisible-tags.

Reference

[idea:config-reference]Argument reference

Parameter	Default	What it sets
`prefix`	`"idea"`	The namespace an idea’s name lives in; non-empty, no `:` (the separator is added for you). See referencing ideas.
`note-dir`	`none`	The directory the minted idea pages are written to. Left unset, it is `ideas` under the default prefix, and the prefix itself under any other — `prefix: "maths"` mints `maths/`, not `mathss/`. Names one directory: no `/` and no `:`.
`css-prefix`	`none`	The stem every class this package emits is built from. Left unset, it follows `prefix`, so renaming the name namespace renames `.idea-box` to `.note-box` along with it; set it to hold the classes still while the names move. See HTML class reference.
`window-unfurl`	`1`	How many levels of transclusion are allowed before a window bottoms out as a link to the idea’s own page. `0` allows none. See controlling window unfurl.
`theme`	`(:)`	The ten keys the package will style for you — five colors, three lengths, a font stack, and `tags-color`. Every key but `tags-color` is also a parameter in its own right, and the granular form wins. See theming a rookery.
`display`	`(:)`	What every idea and window in the rookery shows of itself, as a dictionary of the same nine flags `#idea` takes. This is the bottom of the stack: an idea or a window that leaves a key at auto is asking for the value set here. `date` and `tags` resolve to `false` when nothing is said, the other seven to `true`. Each key is also a parameter in its own right — `display-date: true` — and the granular form wins. See `#idea`.
`idea-page-template`	`none`	Your own chrome for the standalone page rookery mints per idea. See idea template.
`bibliography`	`none`	Typst’s own `#bibliography` arguments, for the single bibliography a rookery’s citations all draw from.
`index-page`	`true`	Whether to mint `ideas/index.html`, a landing page listing every idea in the rookery. It goes in the same directory the idea pages do, so a rookery that moved them with `note-dir` or `prefix` gets `maths/index.html` rather than `ideas/index.html`. Set it to `false` where your site already publishes an index of its own.
`refs`	`true`	Whether rookery installs the `show ref:` rule that renders `@idea:etal` as the idea rather than a figure number. Set it to `false` to keep Typst’s own behavior, or to install a rule of your own. See hyperlinks.
`hyperlink-target-minted`	`true`	Where every `@idea:etal` in the document lands: `true` on the idea’s standalone minted page, `false` in the context it was hatched in. The same parameter `#hyperlink` itself takes. Ignored when `refs` is `false`. See hyperlinks.
`syndicate`	`false`	Whether each minted page also carries a `<feeds:item>` beacon — the label `@rheo/feeds`’s `items()` reads by default — so a feed package can collect your ideas without either package importing the other. Off by default: a package should not write into another’s label namespace unasked. An idea with no date never gets one.
`page-titles`	`"title"`	What the context and backlinks footer on an idea’s page calls the pages it lists: `"title"` is Rheo’s own spine title, `"path"` the source path with the content directory and the extension dropped — `jobs/cfps`, `institutions`. Rheo derives a title from the file stem, so in a project with a page per directory every `index.typ` is titled ‘Index’, which is a poor answer to a footer asking where an idea was written; a path is longer and plainer, and unique by construction. A rookery-wide choice, since one idea’s footer disagreeing with another’s about what a page is called would be worse than either answer.
`invisible-tags`	`()`	Tags that leave no trace in the HTML: no pill in any hat, no `.idea-tag-<tag>` class on the heading, the card, an outline row or an index row, and no generated rule for a `tags-color` entry naming one. Filtering still sees them everywhere — `#window(tagged: ..)`, `#ideas`, `#idea-tag-names`, a search query — which is what makes an invisible tag usable as a build-level key. The presentational complement to `exclude-tags`, which drops the idea itself.