@tomerlichtash/remark-directive-refs

Extract structured refs and anchors from remark-directive attributes — parsing only, no rendering. For aggregate pages (letters, threads) where each item is a directive carrying its own refs:

:::entry{author="Sam Drifter" title="First letter" refs="kit-vale,jo-reed"}
…
:::

import { extractDirectiveRefs, anchorSlug } from '@tomerlichtash/remark-directive-refs';

extractDirectiveRefs(body, {
  directive: 'entry',
  attr: 'refs',                     // comma-separated ids (default 'refs')
  anchorFrom: ['title', 'author'],  // slugged for `anchor`, first non-empty wins
});
// → [{ attrs: { author: 'Sam Drifter', … }, refs: ['kit-vale', 'jo-reed'], anchor: 'First-letter' }]

Bodies are parsed with the same remark-directive your render pipeline uses (not a regex), so attribute quoting/escaping behaves identically. Directives without a non-empty refs attribute are skipped.

anchorSlug(text) is the shared slug rule (whitespace → dash, strip all but letters/numbers/dashes) — import it render-side so backlink #anchors land on the element ids you emit.

Resolving the extracted ids is a separate concern — see @tomerlichtash/md-registry.