code runeAn inline rune that renders its body as code, so a variable can be code-styled. Backticks cannot do it: a code span is literal by definition, so the Markdoc syntax inside one is never parsed.
Tracking started Sep 11 — check back for trends.
ec7357bIt is the escape hatch for the one inline construct that is opaque to Markdoc. Measured inside a {% data %} body, every other inline form passes a variable through intact:
| Authored | Renders |
|---|---|
[{% $row.name %}](/runes/card) | <a href="/runes/card">href</a> |
[**{% $row.name %}**](…) | bold inside a link |
**_{% $row.name %}_** | nested strong + em |
{% badge %}{% $row.name %}{% /badge %} | <span class="rf-badge">href</span> |
`{% $row.name %}` | <code>{% $row.name %}</code> — literal |
badge already proves a rune body takes variables. So backticks are the gap, and a rune is the shape of every other answer in this vocabulary.
Backticks stay the right way to write static code, exactly as **bold** stays the right way to write static emphasis. This rune is for dynamic content, and the docs should say so rather than presenting it as an alternative spelling.
{% code %} renders its body as <code>, and a variable inside it resolvesModel it on badge, the nearest existing inline rune: same posture, same content handling, a different element and no meta fields. Not on snippet or fence — this is an inline span, not a block, and it does no highlighting.
No syntax highlighting. An inline code span is not highlighted today and should not become so here; matching the backtick span exactly is the point.
Completed: 2026-09-11
Branch: claude/content-author-docs-org-vps1un
packages/runes/src/tags/code.ts — the rune. Modelled on badge: inline: true, no attributes, transformChildren into a code element, declareUniversalPosture(code, 'inline').packages/runes/src/index.ts — catalog entry under "Code & Data".packages/runes/src/config.ts — Code: { block: 'code', universalAttributes: 'inline' }.packages/lumina/test/css-coverage.test.ts — added to UNSTYLED_BLOCKS: the rune deliberately renders a bare code element, which global.css's element rule already styles, so there is no .rf-code rule to write.site/content/runes/code.md + nav entry.packages/runes/test/data-body-table.test.ts.The claim that it renders "a bare <code> with no class" was wrong, and the test that "proved" it was measuring the wrong stage. It asserted on Markdoc.transform output, which runs before the identity transform — and the identity transform adds class="rf-code" from the config entry. A real site build showed <code class="rf-code" data-rune="code">. The test now asserts what actually matters and is actually true: the element is code, which is what makes Lumina's code { … } element rule style it identically to a backtick span. .rf-code is a bonus hook for a theme that wants to tell them apart. The rune page was corrected to match.
Scope held. No syntax highlighting, no attributes. An inline backtick span is not highlighted today, and the point of the rune is to be indistinguishable from one.
The docs frame it as the narrow escape hatch it is — backticks stay correct for static code. The rune page carries the measured table showing that links, nested emphasis and rune bodies all pass a variable through, and only a code span does not.