メニュー
REDAXO template authoring – page layouts that wrap article content. Use when the user creates or edits a template, asks about getArticle(), works with rex_fragment, sets up multi-language output, or wants to embed slices into a layout.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Building and maintaining custom REDAXO addons – package.yml, boot.php, install/update/uninstall, lang files, fragments, backend pages, the assets pipeline (and the reinstall-to-sync rule), cache-busting on asset URLs, vendor dependencies, and the common-pitfalls checklist. Use when the user creates a new addon, edits package.yml, sets up backend pages, ships frontend assets from an addon, integrates Composer dependencies, or asks how to package functionality as a reusable addon.
Calling the FriendsOfRedaxo/api REST endpoints (v1.2+) — Bearer + Backend-Session auth, the route table for articles/categories/slices/modules/templates/languages/media/users/metainfo, the slice POST schema, OpenAPI spec, the unified `{data, meta}` list format, and a 401/404/405/500 diagnostic flow. Covers the Backend-mirror under `/api/backend/...`, multipart media upload, metainfo values for articles/categories/media/clangs, and yrewrite cache invalidation. Use when the user calls the api addon over HTTP, hits a confusing 401/404/405, syncs articles between systems via this API, builds tooling around it, or says "API-Aufruf", "REST-API ansprechen", "Bearer-Token", "Slice anlegen per API", "Artikel über API erzeugen".
One-shot recipe to bootstrap a complete YCom community frontend on a REDAXO instance that has the FriendsOfREDAXO `api` addon installed. Builds a clean tree under a Community category — six entry-point sections (login, password-forgot, profile, password-change, registration, logout) as sub-categories so they appear in the meta navigation automatically, and three token-target sub-articles (registration-confirm, password-reset, terms-of-use) tucked under their related section. Wires the YForm pipe code into slices on the standard "YForm Formbuilder" module, sets per-article ycom_auth_type, inserts the email templates, and writes the ycom/auth config keys. Use when the user asks to "set up YCom community pages", "scaffold the YCom frontend", "create the login/registration flow", or runs a fresh REDAXO instance that needs YCom wired up programmatically. Skip when the instance has no `api` addon (offer the manual setup from `ycom-forms` instead).
Search It indexing – generating and managing the search index, reindex articles, index database columns, index files and PDFs, console commands search_it:reindex and search_it:clearCache, cronjobs, plaintext conversion via FriendsOfRedaxo\SearchIt\Plaintext\PlaintextConverter (configured through backend settings + SEARCH_IT_PLAINTEXT extension point) and PDF text extraction via FriendsOfRedaxo\SearchIt\Pdf\PdfConverter (requires pdftotext / poppler-utils on the server). Use when the user asks about rebuilding the search index, indexing custom content, or the search returns stale results.
MultiGlossar replacement engine internals - boot.php OUTPUT_FILTER hook, the (internal) MultiGlossar\\Parser in lib/glossar/multiglossar.php, DOMDocument parsing, text_replace regex matching, replace_definition template placeholders, and URL namespace resolution via Url\\Profile/url_generator_profile. Stable integration is via OUTPUT_FILTER and the addon config — Parser is not a public API. Use when the user debugs missing replacements, broken tooltip HTML, URL addon integration, or replacement performance.
Adding custom REST routes to the FriendsOfRedaxo/api addon (v1.2+) from your own addon — building a RoutePackage subclass, registering routes, scope management, the `Body`/`query` schema keys, request-context caveats (auth runs in frontend context, isBackend() returns false, PRE-EPs that call rex::requireUser() fail under Bearer), and the addon's exact-mirror-of-core convention. Use when the user adds endpoints to the api addon, exposes their own data over the same auth/scope mechanism, runs into "EP fires only in backend but I need it from the API" / "rex::requireUser threw" problems, or says "eigene Route registrieren", "API um eigenen Endpunkt erweitern", "RoutePackage bauen".
| name | redaxo-templates |
| description | REDAXO template authoring – page layouts that wrap article content. Use when the user creates or edits a template, asks about getArticle(), works with rex_fragment, sets up multi-language output, or wants to embed slices into a layout. |
Templates are the page layouts in REDAXO. They include the document shell (<html>, <head>, <body>) plus navigation, header/footer – and call getArticle() to render the article content built from modules.
Each article in the structure addon is assigned a template (with optional inheritance and category-based defaults).
<!DOCTYPE html>
<html lang="<?= rex_clang::getCurrent()->getCode() ?>">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title><?= rex_escape(rex_article::getCurrent()->getName()) ?> – <?= rex_escape(rex::getServerName()) ?></title>
<link rel="stylesheet" href="<?= rex_url::frontend('assets/style.css') ?>">
</head>
<body>
<header>
<a href="<?= rex_getUrl(rex_article::getSiteStartArticleId()) ?>">
<?= rex_escape(rex::getServerName()) ?>
</a>
<?php // Navigation – see "Navigation" below ?>
</header>
<main>
<?php // Render all slices of the current article ?>
<?= REX_ARTICLE[] ?>
</main>
<footer>
<p>© <?= date('Y') ?> <?= rex_escape(rex::getServerName()) ?></p>
</footer>
</body>
</html>
REX_ARTICLE[] (the bracket placeholder) is the standard way to render the current article's slices. There's also REX_ARTICLE[id=5] to render a specific article inline.
REDAXO supports multiple languages via rex_clang. Each language has an ID, a code (e.g. de, en), and a name.
<html lang="<?= rex_clang::getCurrent()->getCode() ?>">
<?php // Language switcher ?>
<ul class="lang-switch">
<?php foreach (rex_clang::getAll(true) as $clang): ?>
<?php $url = rex_getUrl(null, $clang->getId()); ?>
<li class="<?= $clang->getId() === rex_clang::getCurrentId() ? 'active' : '' ?>">
<a href="<?= $url ?>" hreflang="<?= $clang->getCode() ?>">
<?= rex_escape($clang->getName()) ?>
</a>
</li>
<?php endforeach; ?>
</ul>
rex_getUrl(null, $clangId) keeps the current article but switches the language. rex_getUrl($articleId, $clangId) switches both.
The structure addon doesn't ship navigation builders; use one of these:
Manual recursion (small sites):
function renderNav(int $parentId = 0, int $maxDepth = 2, int $depth = 0): string
{
if ($depth >= $maxDepth) {
return '';
}
$items = rex_category::getChildrenById($parentId);
if (!$items) {
return '';
}
$html = '<ul class="nav-level-' . $depth . '">';
foreach ($items as $cat) {
if (!$cat->isOnline()) {
continue;
}
$current = $cat->getId() === rex_category::getCurrentId() ? ' class="active"' : '';
$html .= '<li' . $current . '>';
$html .= '<a href="' . $cat->getUrl() . '">' . rex_escape($cat->getName()) . '</a>';
$html .= renderNav($cat->getId(), $maxDepth, $depth + 1);
$html .= '</li>';
}
return $html . '</ul>';
}
echo renderNav();
Bootstrap-style menus: consider the bootstrap_components or jp_navtagger addons – they generate navigation HTML from the structure.
rex_fragment lets you split templates into reusable pieces. Looks for fragments under <addon>/fragments/ or in your project's fragments/ directory.
$f = new rex_fragment();
$f->setVar('title', $article->getName(), false);
$f->setVar('image', $heroImage);
echo $f->parse('hero.php');
In fragments/hero.php:
<section class="hero">
<h1><?= rex_escape($this->getVar('title')) ?></h1>
<?php if ($img = $this->getVar('image')): ?>
<img src="<?= rex_url::media($img) ?>" alt="">
<?php endif; ?>
</section>
Templates run on every request unless cached. For expensive lookups (DB queries, external APIs), use rex_extension::register('OUTPUT_FILTER', ...) for full-page caching, or store fragments via rex_file::put(rex_path::cache(...)).
rex_article::getSiteStartArticleId().lang attribute on <html> – breaks accessibility and SEO.$_SERVER['HTTP_HOST'] instead of rex::getServer() / rex_yrewrite::getCurrentDomain().rex_article::get() – it returns null.