Wiki Phase 1: categories, drafts/publish, HTML sanitization

Foundation & safety phase of the wiki upgrade (see WIKI_UPGRADE.md).

Schema (additive, idempotent via ensureSchema):
- new wiki_categories table; wiki_pages gains category_id, excerpt,
  published, published_at, sort_order, and a FULLTEXT index
- migration ALTERs guarded with IF NOT EXISTS for existing databases
- seed reworked into 4 sections with the 8 starter pages assigned

Security:
- new utils/sanitizeHtml.js (sanitize-html allowlist); wiki bodies are
  sanitized on every save, and the article renders through DOMPurify
- strips <script>, event handlers (onerror), and javascript: URLs

Backend:
- public: published-only list with ?category filter + /wiki/categories
- admin: extended page CRUD, PATCH publish toggle, category CRUD;
  drafts visible to admin, hidden from public
- all writes logged to activity_log

Frontend:
- data-driven public wiki index (sections + real descriptions; removed
  hardcoded blurbs/Roman numerals) with ?category filtering
- article: category breadcrumb + sanitized render
- admin: Section/Status columns, draft/publish + section + excerpt in the
  editor, and a Manage sections modal

Verified end-to-end against MariaDB 11: migration clean, XSS neutralized,
drafts hidden, client builds, server boots.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
2026-06-27 10:45:21 -05:00
parent dd1f61222d
commit b925114923
20 changed files with 1237 additions and 100 deletions

View File

@@ -0,0 +1,45 @@
const sanitizeHtml = require('sanitize-html')
// Allowlist for wiki/post body HTML. Anything not listed is stripped. This runs
// on every save so the stored value is already safe; the client re-sanitizes on
// render as defense in depth. Tuned for rich-text content from the admin editor.
const OPTIONS = {
allowedTags: [
'h1', 'h2', 'h3', 'h4', 'h5', 'h6',
'p', 'br', 'hr', 'blockquote', 'pre', 'code',
'ul', 'ol', 'li',
'strong', 'b', 'em', 'i', 'u', 's', 'sup', 'sub', 'mark', 'span',
'a', 'img', 'figure', 'figcaption',
'table', 'thead', 'tbody', 'tr', 'th', 'td',
],
allowedAttributes: {
a: ['href', 'name', 'target', 'rel', 'title'],
img: ['src', 'alt', 'title', 'width', 'height'],
span: ['data-wiki-slug'], // marks internal wiki links (used from Phase 3)
th: ['colspan', 'rowspan'],
td: ['colspan', 'rowspan'],
},
// http/https for links and images, mailto for links, plus relative URLs so
// uploaded images (/uploads/...) and internal links (/wiki/...) pass through.
allowedSchemes: ['http', 'https', 'mailto'],
allowedSchemesByTag: { img: ['http', 'https'] },
allowProtocolRelative: false,
// Force safe rel on links that open a new tab; drop empty/odd attributes.
transformTags: {
a: sanitizeHtml.simpleTransform('a', { rel: 'noopener noreferrer nofollow' }, true),
},
disallowedTagsMode: 'discard',
}
/**
* Sanitize a block of body HTML against the allowlist above.
* Null/empty input is returned unchanged.
* @param {string|null|undefined} html
* @returns {string|null|undefined}
*/
function cleanBody(html) {
if (html == null || html === '') return html
return sanitizeHtml(String(html), OPTIONS)
}
module.exports = { cleanBody, OPTIONS }