{"version":"https://jsonfeed.org/version/1.1","title":"Sveleton","home_page_url":"","feed_url":"http://sveltekit-prerender/feed.json","description":"Latest articles from Sveleton","items":[{"id":"/todos/add-og-type-article-meta/","url":"/todos/add-og-type-article-meta/","relPath":"/todos/add-og-type-article-meta/","title":"Add og:type article + article:published_time","summary":"","image":"","tags":["SEO"],"order":{"default":20260515},"authors":["ramil"],"wordCount":111,"date_published":"2026-05-15T00:00:00-00:00","content_html":"
Add og:type article + article:published_time
Ramil Karimovdeveloper
Every todo's PageMeta carries date, updated, and authors, but Meta.svelte:49 always sets og:type to website. None of the article-type signals are emitted, which is\n the largest gap between what the data model promises and what reaches\n crawlers.
Tasks
Detect article-like pages (e.g. presence of date + authors) and set og:type=article
Emit article:published_time from date (ISO via publicationDateToISO)
Emit article:modified_time from updated when set
Emit article:author for each author (resolve via the authors\n map)
Optionally add a per-page meta name=\"author\"
Treat articles as articles
OpenGraph article-type tags from existing PageMeta
There is no static/robots.txt in the repo, and no Sitemap: directive published anywhere. Crawlers default to discovery via common\n paths, which is fragile.
Tasks
Create static/robots.txt with User-agent: *, Allow: /, and a Sitemap: line pointing at PROJECT_URL + '/sitemap-appendix.xml'
"},{"id":"/todos/add-seo-test-coverage/","url":"/todos/add-seo-test-coverage/","relPath":"/todos/add-seo-test-coverage/","title":"Add SEO test coverage","summary":"","image":"","tags":["SEO","Testing"],"order":{"default":20260515},"authors":["ramil"],"wordCount":122,"date_published":"2026-05-15T00:00:00-00:00","content_html":"
Add SEO test coverage
Ramil Karimovdeveloper
The only existing test file is imagetools.test.js. Nothing covers the meta/sitemap/feed\n surface — meaning any of the audit fixes can silently regress without\n notice.
"},{"id":"/todos/centralize-project-constants-in-feeds/","url":"/todos/centralize-project-constants-in-feeds/","relPath":"/todos/centralize-project-constants-in-feeds/","title":"Centralize PROJECT_* in feeds and sitemap","summary":"","image":"","tags":["SEO"],"order":{"default":20260515},"authors":["ramil"],"wordCount":116,"date_published":"2026-05-15T00:00:00-00:00","content_html":"
Centralize PROJECT_* in feeds and sitemap
Ramil Karimovdeveloper
Sveleton is a bootstrap; re-using it for a new project should mean changing PROJECT_NAME, PROJECT_URL, and PROJECT_DESCRIPTION in one place. Today, several SEO outputs bypass these constants.
Current State
feed.json/+server.js:7,17 — hardcodes 'Sveleton' and 'Latest articles from Sveleton'
feed.json/+server.js:15 — home_page_url uses empty base instead of PROJECT_URL
"},{"id":"/todos/filter-sitemap-by-exclusion/","url":"/todos/filter-sitemap-by-exclusion/","relPath":"/todos/filter-sitemap-by-exclusion/","title":"Filter sitemap by excluded/draft/noindex","summary":"","image":"","tags":["SEO"],"order":{"default":0},"authors":["ramil"],"wordCount":111,"date_published":"2026-05-15T00:00:00-00:00","updated":"2026-05-15T00:00:00-00:00","content_html":"
Filter sitemap by excluded/draft/noindex
Ramil Karimovdeveloper
· Updated on
src/routes/sitemap-appendix.xml/+server.js enumerates every +page.svelte from getPages() without honoring the excluded, draft, or noindex flags.\n Engineering todos like fix-ci-type-errors and migrate-to-lightning-css get published to search engines as article content.
Current State
sitemap-appendix.xml/+server.js:46-52 — no filter at all
sitemap/+page.server.js:73 — excludes only the sitemap page itself
pages.remote.js getReadNext:50 — filters both excluded and draft
Tasks
Centralize inclusion policy in a single helper (e.g. isPublic(page))
Honor excluded, draft, noindex
Apply the helper in all three consumers
Stop leaking internal pages
One inclusion policy for sitemap, HTML sitemap, and ReadNext
src/routes/+layout.svelte:28 unconditionally emits <link rel=\"canonical\"> from PROJECT_URL + pathname, while src/lib/Meta.svelte:84-86 also emits one when pageMeta.canonical is set. Pages overriding\n canonical end up with two tags.
Tasks
Remove the unconditional canonical from +layout.svelte
Have Meta.svelte always emit canonical, defaulting to PROJECT_URL + pathname when pageMeta.canonical is not set
"},{"id":"/todos/remove-dead-base-url-and-exclude-paths/","url":"/todos/remove-dead-base-url-and-exclude-paths/","relPath":"/todos/remove-dead-base-url-and-exclude-paths/","title":"Remove dead BASE_URL and EXCLUDE_PATHS","summary":"","image":"","tags":["SEO","Cleanup"],"order":{"default":20260515},"authors":["ramil"],"wordCount":96,"date_published":"2026-05-15T00:00:00-00:00","content_html":"
Remove dead BASE_URL and EXCLUDE_PATHS
Ramil Karimovdeveloper
Three files declare const BASE_URL = '' and use it in string\n concatenations — all no-ops. A separate EXCLUDE_PATHS list points at routes that don't exist in this project (carryover from the\n parent template).
Current State
src/lib/url-utils.js:1
src/lib/page-renderer.svelte.js:8
src/routes/feed.xml/+server.js:23
page-renderer.svelte.js:10-24 — EXCLUDE_PATHS with /blog/, /learning-center/..., /reviews/
Tasks
Delete the three BASE_URL declarations and their usages
Delete EXCLUDE_PATHS or replace with real exclusions
src/routes/+layout.svelte:27 emits <meta name=\"google-site-verification\" content=\"\"> on every page. An empty value does nothing for Search Console and adds noise\n to the head.
Tasks
Either remove the tag, or make it conditional on a real verification\n string (e.g. PROJECT_GOOGLE_VERIFICATION in $lib)
At least 16 demo todo pages carry the identical description: 'Create a demo page for the X component.'. That\n field is what Meta.svelte emits as <meta name=\"description\">, so search engines see duplicate\n descriptions across the site.
Recent commits replaced NavAnchor with the navSection attachment and renamed PublicationAuthors to Authors, but the project's CLAUDE.md hasn't caught up.
Tasks
Replace NavAnchor usage in the todo page template with <h2 id=\"...\" {@attach navSection()}>
Update the exports list to remove NavAnchor, PublicationAuthors, and pages; add Authors
Document that Meta auto-resolves from PageMeta — todo pages should not add a per-page <Meta>
Document the SEO outputs: sitemap-appendix.xml, feed.json, feed.xml
Fix README's Node version (24) and the imagetools-disabled claim
Import each image explicitly. Useful when you need a specific subset or want\n full control over order.
<script>\nimport { Gallery } from '$lib';\nimport hero from './assets/hero.jpg';\nimport detail from './assets/detail.jpg';\nimport wide from './assets/wide.jpg';\n</script>\n\n<Gallery images={[hero, detail, wide]} />
Mixed sources
Combine glob and individual imports. For example, glob a folder and prepend\n a cover image.
Pass an optional captions prop as a Record<string, string> keyed by the same paths used in the images record. Captions\n appear as <figcaption> in both the grid and the dialog.
"},{"id":"/todos/move-deploy-to-ci/","url":"/todos/move-deploy-to-ci/","relPath":"/todos/move-deploy-to-ci/","title":"Move Deploy to CI","summary":"","image":"","tags":["CI","Deploy"],"order":{"default":0},"authors":["ramil"],"wordCount":79,"date_published":"2026-04-13T00:00:00-00:00","updated":"2026-04-17T00:00:00-00:00","content_html":"
Move Deploy to CI
Ramil Karimovdeveloper
· Updated on
Currently deployment is manual via ./deploy.sh. Move it to\n GitHub Actions so every push to main deploys automatically to sveleton.ramilkarimov.me.
Tasks
Add SSH_PRIVATE_KEY secret to GitHub repo
Add deploy job to .github/workflows/ci.yml (runs after build,\n on main only)
Deploy via rsync to root@universal.ramilkarimov.me:/var/www/sveleton
"},{"id":"/todos/add-test-run-to-ci/","url":"/todos/add-test-run-to-ci/","relPath":"/todos/add-test-run-to-ci/","title":"Add test run to CI","summary":"","image":"","tags":["Testing","CI"],"order":{"default":0},"authors":["ramil"],"wordCount":67,"date_published":"2026-04-06T00:00:00-00:00","updated":"2026-04-17T00:00:00-00:00","content_html":"
Add test run to CI
Ramil Karimovdeveloper
· Updated on
pnpm test:ci (vitest run) is set up and passing locally. Add it to .github/workflows/ci.yml before the build step so failures are caught early.
Change
In .github/workflows/ci.yml, add after pnpm check:
"},{"id":"/reports/audit-meta-refactoring/","url":"/reports/audit-meta-refactoring/","relPath":"/reports/audit-meta-refactoring/","title":"Audit: Meta Refactoring","summary":"","image":"","tags":["Audit","SEO"],"order":{"default":0},"authors":["ramil"],"excluded":true,"wordCount":798,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Audit: Meta Refactoring
Ramil Karimovdeveloper
Comprehensive audit of the Meta.svelte self-resolving refactoring, performed\n by 8 specialized agents in parallel.
Executive Summary
The refactoring successfully makes Meta.svelte self-resolve\n title, description, and new SEO tags from PageMeta exports via\n synchronous import.meta.glob. The build succeeds and meta tags render\n correctly in both dev and production. However, the refactoring is\n incomplete: 66 page files still pass a meta prop to <Publication> that no longer accepts it, and CLAUDE.md still documents the old API.
Critical Issues
1. Orphaned meta prop on 66 Publication calls
Publication.svelte no longer accepts a meta prop, but 66 page files still pass meta={{ title, description }}. Svelte\n silently ignores unknown props at runtime, so there is no error, but this is\n dead code that creates API confusion.
2. CLAUDE.md shows outdated Publication API
Lines 82-90 of CLAUDE.md show the old pattern with meta={{ title, description }} on <Publication>. This no longer works since Publication\n dropped the prop.
3. Unused SharePostLink import in PublicationHead.svelte
PublicationHead.svelte line 6 imports SharePostLink from $lib but never uses it. This was likely left over from a previous version.
Maintainability Concerns
Pages map built 3 times
The same import.meta.glob('/**/+page.svelte', { import: 'meta', eager: true\n }) pattern is duplicated in 3 files:
Meta.svelte (for head tag resolution)
Articles.svelte (for article listings)
pages.remote.js (for server-side metadata)
Each builds a nearly identical pages map with the same pathToPublicURL + filter + cover-strip logic. If the glob pattern or mapping logic needs to\n change, all 3 must be updated.
getCoverUrl duplicated
The getCoverUrl(cover) helper exists in both Meta.svelte and page-renderer.svelte.js with identical logic.
Magic numbers for OG image dimensions
Meta.svelte hardcodes 1200 and 630 for OG image width and height. These are standard Open Graph sizes but would\n be clearer as named constants.
Consistency Audit
Glob pattern: Consistent
All 3 files use the same glob pattern, pathToPublicURL prefix ( '/src/routes/'), and cover-stripping destructure. The pattern\n matches the established convention in Articles.svelte.
Indentation: Mixed
Some files use tabs (newer todo pages), others use spaces (older pages, lib\n components). This is a pre-existing inconsistency not introduced by this\n refactoring.
Title variations: By design
Some pages have different titles in PageMeta.title vs the\n former Publication meta.title (e.g., \"Sync editor settings\" vs \"Sync Biome, Zed and VS Code settings\").\n Since Meta now resolves from PageMeta.title, the shorter title will appear in <title>. This may or may not be desired.
Dead Code
SharePostLink import in PublicationHead.svelte —\n imported but unused
66 meta={{...}} props on Publication\n calls — passed but ignored
YandexMetricsEventShareLink in SharePostLink.svelte — defined but never called (pre-existing)
AI Artifacts
Overly verbose JSDoc comments exist in several files (notably Articles.svelte with a 40-line file-level comment block). These are pre-existing and not\n introduced by this refactoring. The new code in Meta.svelte has\n appropriate-level documentation.
Test Coverage
No tests exist. Vitest 4.0.18 is installed as a devDependency but no test\n script, config, or test files exist. Key functions that should be tested:
pathToPublicURL — core routing logic, pure function, easy to test
"},{"id":"/todos/add-feed-link-tags/","url":"/todos/add-feed-link-tags/","relPath":"/todos/add-feed-link-tags/","title":"Add feed link tags to layout","summary":"","image":"","tags":["Features"],"order":{"default":0},"authors":["ramil"],"wordCount":104,"date_published":"2026-02-09T00:00:00-00:00","updated":"2026-05-04T00:00:00-00:00","content_html":"
Add feed link tags to layout
Ramil Karimovdeveloper
· Updated on
The feed link tags in +layout.svelte were commented out during\n initial setup and never restored. Without <link rel=\"alternate\"> tags, feed readers and browsers cannot auto-discover the site's JSON and XML\n feeds.
Current State
feed.json and feed.xml routes are restored and working
+layout.svelte has no <link rel=\"alternate\"> tags pointing to either feed
Tasks
Add <link rel=\"alternate\" type=\"application/feed+json\" href=\"/feed.json\"> to <svelte:head> in +layout.svelte
Add <link rel=\"alternate\" type=\"application/atom+xml\" href=\"/feed.xml\"> to <svelte:head> in +layout.svelte
"},{"id":"/todos/add-fields-to-page-meta/","url":"/todos/add-fields-to-page-meta/","relPath":"/todos/add-fields-to-page-meta/","title":"Add Fields to PageMeta","summary":"","image":"","tags":["Types"],"order":{"default":0},"authors":["ramil"],"wordCount":220,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Add Fields to PageMeta
Ramil Karimovdeveloper
The PageMeta interface in app.d.ts defines the\n metadata structure for all pages. It currently supports title, date,\n authors, tags, sort order, cover image, and todo state. This task is about\n extending it with additional fields to support richer page metadata.
Current Fields
title — page title (required)
menuTitle — alternative title for navigation menus
date — publication date in DD-MM-YYYY format
updated — last updated date
authors — author tags or full Author objects
tags — content tags
sortOrder — numeric sort order for listings
cover — cover image
state — todo state (done, new, in_progress)
Candidate Fields
description — SEO meta description, currently only passed to Publication
draft — boolean to mark unpublished pages
featured — boolean or number for featured/pinned content
readingTime — estimated reading time override
canonical — canonical URL for cross-posted content
noindex — exclude page from search engines
lang — page language if different from site default
category — broader content category beyond tags
Tasks
Decide which fields to add to PageMeta
Update the interface in app.d.ts
Update components that consume PageMeta to use new fields
"},{"id":"/todos/add-footer-links/","url":"/todos/add-footer-links/","relPath":"/todos/add-footer-links/","title":"Add links to footer","summary":"","image":"","tags":["UI"],"order":{"default":0},"authors":["ramil"],"wordCount":60,"date_published":"2026-02-09T00:00:00-00:00","updated":"2026-02-09T00:00:00-00:00","content_html":"
Add links to footer
Ramil Karimovdeveloper
· Updated on
Add links to the glossary page and the sitemap into the footer for better\n site navigation.
"},{"id":"/todos/add-tests/","url":"/todos/add-tests/","relPath":"/todos/add-tests/","title":"Add tests and CI integration","summary":"","image":"","tags":["Testing","CI"],"order":{"default":20260209},"authors":["ramil"],"featured":true,"wordCount":184,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Add tests and CI integration
Ramil Karimovdeveloper
Vitest is already installed (^4.0.18) but has no configuration,\n no test scripts, and no test files. This is the umbrella task, split into 4\n sub-tasks that should be done in order.
Component tests -- Svelte component rendering with @testing-library/svelte
Execution order
Task 1 (environment setup) must be done first. Tasks 2 and 3 can be done in\n parallel after that. Task 4 (component tests) depends on task 1 and may\n require additional dependencies (@testing-library/svelte).
E2E tests (future consideration)
Playwright for full browser testing. Lower priority since this is a static\n site, but useful for interactive components.
After adding new fields to PageMeta, several inconsistencies\n remain between the interface, its consumers, and the runtime data flow. This\n task addresses duplication, missing type connections, and stale patterns.
Issues
description is defined in PageMeta but not wired through — pages still\n duplicate it in the Publication meta prop
page-renderer.svelte.js reads pageMetadata as untyped Record<string, unknown> — should use PageMeta
FeedItem in app.d.ts duplicates fields from PageMeta (tags, authors, cover) with slightly different types
Publication.svelte MetaConfig typedef duplicates fields that could come from PageMeta
PageModule interface (title, menuTitle) overlaps with PageMeta — consider removing it
The cover field is typed as unknown in PageMeta but has a specific shape in FeedItem
Tasks
Wire PageMeta.description through Publication so pages don't\n need to pass it twice
Type pageMetadata in page-renderer.svelte.js as PageMeta
Align FeedItem cover type with PageMeta cover\n type
Consider deriving FeedItem fields from PageMeta using Pick
Svelte 5 component tests to verify rendering output and behavior. Uses @testing-library/svelte for DOM-based tests or Svelte's render from svelte/server for SSR snapshot tests.
"},{"id":"/todos/fix-ci-type-errors/","url":"/todos/fix-ci-type-errors/","relPath":"/todos/fix-ci-type-errors/","title":"Fix CI type errors and lint errors","summary":"","image":"","tags":["Cleanup"],"order":{"default":20260209},"authors":["ramil"],"wordCount":79,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Fix CI type errors and lint errors
Ramil Karimovdeveloper
Both pnpm check and pnpm lint are currently\n failing, which means CI is broken.
Type Errors
excluded and featured properties missing from PageMeta interface in src/app.d.ts — added
excluded missing from FeedItem interface — added
cover variable used before declaration in ~30 todo pages
Dev server logs show: [404] GET /todos/add-fields-to-page-meta/installHook.js.map
This is likely a browser extension (React DevTools, Vue DevTools, etc.)\n injecting installHook.js into the page and then requesting its source map relative to the current\n page URL. Need to confirm the source and decide whether to suppress or fix.
Things to Check
Does it reproduce with all browser extensions disabled?
Which extension injects installHook.js?
Does it happen on all todo pages or only this one?
Is it harmless noise or does it affect the build/SSG output?
"},{"id":"/todos/move-meta-to-root-layout/","url":"/todos/move-meta-to-root-layout/","relPath":"/todos/move-meta-to-root-layout/","title":"Move Meta tag into root layout","summary":"","image":"","tags":["Architecture"],"order":{"default":0},"authors":["ramil"],"wordCount":244,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Move Meta tag into root layout
Ramil Karimovdeveloper
Currently Meta.svelte is imported and rendered manually in\n individual pages, listing pages, and inside PublicationHead.svelte. Since Meta already self-resolves\n title/description/tags from PageMeta via the current URL pathname, it can\n live in the root layout and work for every page automatically.
Current usage
Meta is imported and used in these places:
+page.svelte -- homepage
+error.svelte -- error page
todos/+page.svelte -- todos listing
features/+page.svelte -- features listing
PublicationHead.svelte -- all publications
Several standalone todo pages
Goal
Add <Meta /> to src/routes/+layout.svelte with no props. It will auto-resolve from PageMeta for every page. Then\n remove all the individual <Meta> usages.
Steps
Add Meta to +layout.svelte
Remove Meta from PublicationHead
Remove Meta from homepage, todos, features
Remove Meta from standalone todo pages
Handle the error page -- it uses explicit props since error pages don't\n have PageMeta. Keep Meta there as a fallback, or add error-specific logic\n to layout Meta.
Verify all pages still have correct meta tags
Edge cases
Error page: passes explicit title/description props. Meta falls back to\n props when PageMeta is not found, so it could stay in both error page and\n layout. Check for duplicate title tags.
Homepage: passes explicit props. Verify PageMeta has the correct\n title/description.
Duplicate canonical: root layout already renders canonical link. Meta also\n renders one when pageMeta.canonical is set. Avoid conflicts.
"},{"id":"/todos/organize-distribution-options/","url":"/todos/organize-distribution-options/","relPath":"/todos/organize-distribution-options/","title":"Organize Distribution Options","summary":"","image":"","tags":["Infrastructure"],"order":{"default":20260209},"authors":["ramil"],"wordCount":136,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Organize Distribution Options
Ramil Karimovdeveloper
Explore and set up distribution options so the project can be shipped as a\n static bundle, Docker image, or deployed to edge functions depending on the\n target environment.
Options to evaluate
Static build — current default via @sveltejs/adapter-static, outputs to build/
Docker — containerized build for self-hosting (Nginx/Caddy serving static files,\n or Node server)
"},{"id":"/todos/replace-navanchor-with-anchors/","url":"/todos/replace-navanchor-with-anchors/","relPath":"/todos/replace-navanchor-with-anchors/","title":"Replace NavAnchor with normal anchors","summary":"","image":"","tags":["Architecture"],"order":{"default":0},"authors":["ramil"],"wordCount":274,"date_published":"2026-02-09T00:00:00-00:00","updated":"2026-05-04T00:00:00-00:00","content_html":"
Replace NavAnchor with normal anchors
Ramil Karimovdeveloper
· Updated on
The NavAnchor component is used throughout publication pages to\n create section anchors for the sidebar navigation. It could be replaced with\n plain HTML anchor elements, reducing the component abstraction.
Current behavior
NavAnchor does three things:
Renders a <div> with an id attribute for\n anchor linking
Registers itself with the sidebar Nav component via addItem()/removeItem() (from Nav.svelte)
Tracks its DOM position via use:updateTopPosition action\n for scroll-spy behavior
The key complexity is that NavAnchor is not just an anchor -- it registers\n with the sidebar Nav for table-of-contents generation and scroll-spy.\n Removing NavAnchor requires an alternative way to:
Discover section headings for the sidebar table of contents
Track scroll position to highlight the active section
Possible approaches
Option A: DOM-based heading discovery
Nav component queries the DOM for headings with id attributes\n (e.g. h2[id]) and builds the TOC automatically. Uses IntersectionObserver for scroll-spy instead of manual position tracking.
Option B: Svelte action on headings
Create a use:navSection action that replaces NavAnchor's\n registration logic. Headings become <h2 id=\"foo\" use:navSection>.
Option C: Keep registration, lose the component
Move the anchor+registration into an action or into the heading elements\n directly, but still use the addItem/removeItem pattern from Nav.
Scope
NavAnchor is used in many pages across the codebase. This is a bulk\n replacement task.
All publication pages (via Publication wrapper pattern)
"},{"id":"/todos/restore-og-image-generation/","url":"/todos/restore-og-image-generation/","relPath":"/todos/restore-og-image-generation/","title":"Restore dynamic OG image generation","summary":"","image":"","tags":["Features"],"order":{"default":20260209},"authors":["ramil"],"wordCount":142,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Restore dynamic OG image generation
Ramil Karimovdeveloper
The og/ route was disabled during the initial setup. Currently Meta.svelte uses the cover image directly via getCoverUrl() instead of a purpose-built 1200x630 OG image. Social sharing previews are\n either missing or use an arbitrarily sized cover.
Current State
No src/routes/og/ route exists
Meta.svelte sets og:image to the raw cover image src (getCoverUrl(cover))
OG image meta tags claim 1200x630 PNG but the actual image may be a\n different size and format
Tasks
Create src/routes/og/[...path]/+server.js route that\n generates 1200x630 OG images
Decide on generation approach (satori, puppeteer, canvas, or static\n template)
Update Meta.svelte to point og:image at the\n generated OG endpoint instead of the raw cover
Ensure OG images are prerendered at build time for static hosting
"},{"id":"/todos/setup-testing-environment/","url":"/todos/setup-testing-environment/","relPath":"/todos/setup-testing-environment/","title":"Set up testing environment and CI","summary":"","image":"","tags":["Testing","CI"],"order":{"default":0},"authors":["ramil"],"wordCount":222,"date_published":"2026-02-09T00:00:00-00:00","updated":"2026-05-15T00:00:00-00:00","content_html":"
Set up testing environment and CI
Ramil Karimovdeveloper
· Updated on
Vitest ^4.0.18 is installed but completely unconfigured. This\n task sets up the foundation so that subsequent test-writing tasks can just\n add .test.js files and they work.
"},{"id":"/todos/unit-tests-endpoints/","url":"/todos/unit-tests-endpoints/","relPath":"/todos/unit-tests-endpoints/","title":"Unit tests for endpoints and renderer","summary":"","image":"","tags":["Testing"],"order":{"default":20260207},"authors":["ramil"],"wordCount":377,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Unit tests for endpoints and renderer
Ramil Karimovdeveloper
Tests for server-side logic: sitemap generation helpers, RSS feed sanitizer,\n and the page renderer. Most of these functions are currently not exported\n and need to be made testable first.
"},{"id":"/todos/unit-tests-utilities/","url":"/todos/unit-tests-utilities/","relPath":"/todos/unit-tests-utilities/","title":"Unit tests for utility functions","summary":"","image":"","tags":["Testing"],"order":{"default":20260208},"authors":["ramil"],"wordCount":291,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Unit tests for utility functions
Ramil Karimovdeveloper
Pure functions that can be tested directly with minimal or no mocking. These\n are the easiest tests to write and provide immediate value.
"},{"id":"/todos/update-meta-component/","url":"/todos/update-meta-component/","relPath":"/todos/update-meta-component/","title":"Update Meta Component","summary":"","image":"","tags":["SEO"],"order":{"default":0},"authors":["ramil"],"wordCount":160,"date_published":"2026-02-09T00:00:00-00:00","updated":"2026-02-09T00:00:00-00:00","content_html":"
Update Meta Component
Ramil Karimovdeveloper
· Updated on
Meta.svelte now self-resolves title, description, and SEO tags using a synchronous import.meta.glob PageMeta lookup keyed by the current pathname via $app/state.\n Props are still accepted as fallbacks for pages without PageMeta exports (error page).
What Changed
Meta.svelte
Builds a pages map via import.meta.glob and resolves pages[page.url.pathname]. Title and description come from PageMeta with prop fallbacks. New tags added:
<link rel=\"canonical\"> from canonical
<meta name=\"robots\"> from noindex
<meta property=\"og:locale\"> from lang
<meta name=\"keywords\"> from tags
Publication / PublicationHead
meta prop removed from Publication and PublicationHead. <Meta /> is now called with\n no props.
Page meta exports
Added description field to every page's PageMeta export, migrated from the former Publication meta prop.
"},{"id":"/todos/use-img-component-for-author-avatars/","url":"/todos/use-img-component-for-author-avatars/","relPath":"/todos/use-img-component-for-author-avatars/","title":"Use Img component for author avatars","summary":"","image":"","tags":["Features"],"order":{"default":0},"authors":["ramil"],"wordCount":132,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Use Img component for author avatars
Ramil Karimovdeveloper
Author avatars are stored in static/images/authors/ and the avatar field is defined in src/lib/authors.js, but PublicationAuthors.svelte doesn't render them yet. Use the common Img component so\n avatars go through the same image pipeline as everything else (srcset,\n LQIP).
Current State
Author interface in src/app.d.ts has an optional avatar field
src/lib/authors.js has avatar paths set (e.g. /images/authors/ramil.avif)
PublicationAuthors.svelte only renders author names, no images
Tasks
Update PublicationAuthors.svelte to render avatar via Img component
Look up full author data from authors map using the tag
Handle missing avatars gracefully (initials fallback or hide image)
Migrate from export let to $props() while at it\n (Svelte 5)
"},{"id":"/todos/use-remote-function-to-read-next/","url":"/todos/use-remote-function-to-read-next/","relPath":"/todos/use-remote-function-to-read-next/","title":"Use remote function to ReadNext","summary":"","image":"","tags":["Content"],"order":{"default":0},"authors":["ramil"],"wordCount":141,"date_published":"2026-02-09T00:00:00-00:00","content_html":"
Use remote function to ReadNext
Ramil Karimovdeveloper
The ReadNext component should suggest related articles based on\n shared tags, excluding the current article.
Current State
Currently reads from page.data.readNextPosts which is not\n populated. The component needs to self-resolve suggestions from the feed.
Requirements
Load articles from the feed that share at least one tag with the current\n article
Exclude the current article from results
Sort by number of shared tags (most relevant first)
Limit to 3-4 suggestions
Fall back to recent articles if no tag matches found
Tasks
Resolve current page tags from PageMeta (via buildPagesMap or\n feed data)
Filter feed items by shared tags, excluding current URL
Sveleton provides a collection of reusable UI components optimized for\n content-rich media projects. All components are exported from $lib for easy importing throughout your application.
Key Components
The component library includes the following building blocks:
Articles - Grid layout component for displaying collections of articles or posts.
PostPreview - Card component for article previews with metadata and thumbnails.
Accordion - Collapsible content sections for FAQ-style layouts.
Img - Responsive image component powered by @zerodevx/svelte-img with automatic srcset generation and LQIP placeholders.
LinkBlock - Clickable card component for featured links and CTAs.
SharePostLink - Social sharing buttons for content distribution.
CallToAction - Prominent CTA sections for conversion optimization.
AnchorHeading - Headings with automatic anchor links for deep linking.
"},{"id":"/features/feeds/","url":"/features/feeds/","relPath":"/features/feeds/","title":"JSON and XML Feeds","summary":"","image":"/_app/immutable/assets/cover.lbay0dt-.avif","tags":["Content"],"order":{"default":20260208},"authors":["ramil"],"wordCount":154,"date_published":"2026-02-08T00:00:00-00:00","content_html":"
JSON and XML Feeds
Ramil Karimovdeveloper
Sveleton automatically generates content feeds for RSS readers and\n aggregators. Both modern JSON Feed and traditional Atom XML formats are\n supported, ensuring your content reaches the widest possible audience.
JSON Feed
The /feed.json endpoint serves a JSON Feed 1.1 compliant feed.\n This modern format is easier to parse and more developer-friendly than XML.\n The feed includes all your published articles with titles, descriptions,\n dates, authors, and cover images.
Content extraction happens server-side using Cheerio to parse each article's\n HTML and generate clean feed items. This ensures feed readers get properly\n formatted content without any layout artifacts.
XML Feed
The /feed.xml endpoint provides an Atom-format XML feed for\n traditional RSS readers. Like the JSON feed, it includes full article\n metadata and content, ensuring compatibility with all major feed readers and\n aggregation services.
Sveleton automatically optimizes images at build time using vite-imagetools. Every image is transformed into multiple sizes\n and formats, ensuring fast loading times across all devices and network\n conditions.
How It Works
Images are processed during the Vite build step, generating responsive\n variants from 197px to 1920px wide in AVIF format. The system automatically\n creates a srcset attribute with all variants, letting browsers\n choose the optimal size for each viewport.
LQIP (Low Quality Image Placeholder) support provides instant visual\n feedback. A tiny, blurred version of the image appears immediately while the\n full image loads in the background, improving perceived performance.
Usage
Import images directly — no query parameters needed. Responsive variants and\n LQIP placeholders are applied automatically via defaultDirectives in vite.config.js.
The Nav component provides an automatic sidebar table of\n contents that generates navigation links from NavAnchor elements placed throughout your content. Simply add NavAnchor components before each section heading to create jump links.
The navigation updates automatically as users scroll, highlighting the\n current section and providing smooth anchor-based scrolling between content\n sections.
Breadcrumbs
The Breadcrumbs component generates hierarchical navigation\n automatically based on your route structure. It helps users understand their\n current location within the site and provides quick access to parent pages.
For site-wide navigation, the sitemap page provides a complete\n overview of all available content organized by category and type.
The Publication component is the foundation of Sveleton's\n article publishing system. It wraps your content and automatically adds all\n the metadata, navigation, and layout elements needed for a professional\n publication.
Overview
Every article wrapped in the Publication component gets\n automatic breadcrumb navigation, author bylines, publication and update\n dates, and estimated reading time. The component also handles the article's\n table of contents navigation, allowing readers to jump between sections\n using NavAnchor components.
At the end of each article, the ReadNext carousel automatically\n suggests related content based on tags and categories. This keeps readers\n engaged and helps them discover more of your content.
Components
The system includes several supporting components: PublicationAuthors for displaying author information with avatars and links, Breadcrumbs for site navigation, Nav for the in-page table of contents, and CallToAction for adding engagement prompts at the end of articles.
"},{"id":"/features/seo-and-meta-tags/","url":"/features/seo-and-meta-tags/","relPath":"/features/seo-and-meta-tags/","title":"SEO and Meta Tags","summary":"","image":"/_app/immutable/assets/cover.C3Fpq6Lk.avif","tags":["Content"],"order":{"default":20260208},"authors":["ramil"],"wordCount":182,"date_published":"2026-02-08T00:00:00-00:00","content_html":"
SEO and Meta Tags
Ramil Karimovdeveloper
Sveleton includes a comprehensive Meta component that handles\n all your SEO needs out of the box. Every page gets proper meta tags for\n search engines and social media platforms without any additional\n configuration.
What Is Included
The Meta component automatically generates OpenGraph tags for\n Facebook and LinkedIn, Twitter Card markup for enhanced tweet previews, and\n Schema.org Article structured data for rich search results. All pages\n include canonical URLs to prevent duplicate content issues.
Each publication page can specify its own title, description, and cover\n image through the meta prop. These values are used consistently\n across all meta tag formats to ensure your content appears correctly when\n shared on any platform.
Usage
Simply wrap your content in the Publication component and pass\n the meta object with your page's title, description, and optional cover\n image. The component handles everything else automatically, including\n generating the proper meta tags, setting up breadcrumbs, and injecting\n Schema.org markup into the page head.
"},{"id":"/features/static-site-generation/","url":"/features/static-site-generation/","relPath":"/features/static-site-generation/","title":"Static Site Generation","summary":"","image":"/_app/immutable/assets/cover.CR8QPPm9.avif","tags":["Build"],"order":{"default":20260208},"authors":["ramil"],"wordCount":185,"date_published":"2026-02-08T00:00:00-00:00","content_html":"
Static Site Generation
Ramil Karimovdeveloper
Sveleton uses SvelteKit's adapter-static to generate fully\n pre-rendered HTML pages at build time. Every route is converted to static\n HTML and output to the build/ directory, making the site\n incredibly fast and suitable for hosting on any static file server.
How It Works
During the build process, SvelteKit crawls all routes starting from the root\n and generates static HTML files. The adapter outputs a 200.html fallback file for client-side routing, ensuring smooth navigation between\n pages without full page reloads.
All pages are pre-rendered by default using SvelteKit's prerender = true configuration. This means content is generated once at build time, not on\n each request, resulting in optimal performance and SEO.
Benefits
Static site generation offers significant advantages for media projects:\n blazing-fast page loads, perfect SEO crawlability, and simple deployment to\n any CDN or static host. No server required.
"},{"id":"/todos/migrate-to-lightning-css/","url":"/todos/migrate-to-lightning-css/","relPath":"/todos/migrate-to-lightning-css/","title":"Migrate to Lightning CSS","summary":"","image":"","tags":["Build"],"order":{"default":0},"authors":["ramil"],"wordCount":174,"date_published":"2026-02-08T00:00:00-00:00","content_html":"
Migrate to Lightning CSS
Ramil Karimovdeveloper
Replace the current PostCSS pipeline with Lightning CSS. Lightning CSS is a\n fast CSS parser, transformer, and minifier written in Rust, with built-in\n support for nesting, custom media queries, and modern CSS features without\n needing separate plugins.
Motivation
Significantly faster CSS processing (Rust-based)
Native support for CSS nesting, no plugin needed
Built-in vendor prefixing and browser target support
Vite has built-in Lightning CSS integration
Reduce dependency count by removing PostCSS plugins
Tasks
Enable css.lightningcss in vite.config.js
Replace postcss-nested-props usage with standard CSS nesting
Replace postcss-simple-vars with CSS custom properties
Replace postcss-easy-import with native CSS imports
"},{"id":"/todos/migrate-lighthouse-and-links-check/","url":"/todos/migrate-lighthouse-and-links-check/","relPath":"/todos/migrate-lighthouse-and-links-check/","title":"Migrate Lighthouse and Links Check","summary":"","image":"","tags":["Infrastructure"],"order":{"default":20260206},"authors":["ramil"],"wordCount":184,"date_published":"2026-02-06T00:00:00-00:00","content_html":"
Migrate Lighthouse and Links Check
Ramil Karimovdeveloper
Migrate Lighthouse CI audits and broken links checking into this repository.
Context
SvelteKit's prerender crawler only validates internal links. External URLs (https://...) are silently discarded by is_root_relative() in the prerender loop — no handler, callback, or plugin hook ever sees them.\n There is no built-in or planned feature for external link validation.
Approach
Use lychee (same as\n timestripe) as a post-build link checker against the build/ output. Lychee is a fast Rust-based tool that checks both internal and\n external links in HTML files.
Install via brew install lychee (local) and GitHub Action\n (CI)
Run against ./build/ after pnpm build
Add a check:links script (not postbuild — avoid\n slowing local dev)
Configure .lychee.toml for excludes, timeouts, accepted\n status codes
Tasks
Migrate Lighthouse CI configuration and scripts
Add lychee as the broken links checker (replaces previous\n solution)
Add .lychee.toml config with sensible defaults
Add check:links script to package.json
Integrate both into the CI/CD pipeline (GitHub Actions)
"},{"id":"/todos/rename-project-to-sveleton/","url":"/todos/rename-project-to-sveleton/","relPath":"/todos/rename-project-to-sveleton/","title":"Rename Project to Sveleton","summary":"","image":"","tags":["Project"],"order":{"default":0},"authors":["ramil"],"wordCount":52,"date_published":"2026-02-02T00:00:00-00:00","content_html":"
Rename Project to Sveleton
Ramil Karimovdeveloper
Rename the project from Sveleton to Sveleton for a cleaner, shorter name.
Create a demo page showcasing the ReadNext component, which\n suggests related articles based on shared tags.
Current State
The ReadNext implementation is complete (see Use remote function to ReadNext). It reads from page.data.readNextPosts populated by getReadNext() in +layout.server.js, and delegates rendering to Articles in compact mode.
Tasks
Create a demo page with usage examples
Show how tag-based matching and fallback behavior works
"},{"id":"/todos/fix-time-to-read/","url":"/todos/fix-time-to-read/","relPath":"/todos/fix-time-to-read/","title":"Fix time to read","summary":"","image":"","tags":["Features"],"order":{"default":0},"authors":["ramil"],"wordCount":59,"date_published":"2026-01-27T00:00:00-00:00","updated":"2026-02-09T00:00:00-00:00","content_html":"
Fix time to read
Ramil Karimovdeveloper
· Updated on
Fix the reading time estimation feature to accurately calculate and display\n reading time for articles.
"},{"id":"/daler/","url":"/daler/","relPath":"/daler/","title":"Todos for Daler","summary":"","image":"","order":{"default":20260126},"wordCount":3,"date_published":"2026-01-26T00:00:00-00:00","content_html":"
Todos for Daler
"},{"id":"/todos/add-optional-snippet-for-article-preview/","url":"/todos/add-optional-snippet-for-article-preview/","relPath":"/todos/add-optional-snippet-for-article-preview/","title":"Add optional snippet for article preview","summary":"","image":"","tags":["Features"],"order":{"default":0},"authors":["ramil"],"wordCount":191,"date_published":"2026-01-26T00:00:00-00:00","updated":"2026-02-09T00:00:00-00:00","content_html":"
Add optional snippet for article preview
Ramil Karimovdeveloper
· Updated on
Currently the PostPreview component only displays the article\n title, cover image, and authors. Adding an optional Svelte snippet would\n allow rich preview content for articles.
Problem
Svelte snippets cannot be passed through JSON. The current architecture:
Pages export metadata in <script module>
page-renderer.svelte.js extracts exports into JSON feed
Articles.svelte reads from page.data.feed (JSON)
PostPreview.svelte renders from JSON data
Svelte snippets are runtime constructs that cannot be serialized to JSON.
Solution
Import all page modules directly in the layout and map them to feed items,\n bypassing JSON serialization for snippet data.
Tasks
Import all page modules in layout using import.meta.glob
Create a map from feed item URLs to page modules
Extract snippet from page module and pass to PostPreview
Update PostPreview.svelte to accept and render optional\n snippet prop
Considerations
Performance impact of importing all page modules in layout
May want different snippet visibility for compact vs fancy layouts
Snippet should be optional - fall back to title-only preview
"},{"id":"/todos/create-new-homepage/","url":"/todos/create-new-homepage/","relPath":"/todos/create-new-homepage/","title":"Create new homepage","summary":"","image":"","tags":["Content"],"order":{"default":0},"authors":["ramil"],"wordCount":126,"date_published":"2026-01-26T00:00:00-00:00","content_html":"
Create new homepage
Ramil Karimovdeveloper
Currently the homepage redirects to /todos/. Create a proper\n landing page for the Sveleton project.
About Sveleton
Sveleton is a bootstrap project based on the Intuition Design System, built\n on the foundations of intuition.team and timestripe.com for quick-starting\n new media projects with everything ready for content writing, SEO, and more.
Requirements
Welcome message explaining what Sveleton is (see above)
Links to key sections (Todos, Sitemap)
Clean, minimal design consistent with the rest of the site
Proper meta tags for SEO
Implementation
Replace src/routes/+page.server.js redirect with actual +page.svelte
Create a pages metadata module using a SvelteKit remote function (prerender) instead of a plain client-side module with import.meta.glob. This keeps the glob/metadata resolution on\n the server and prerendered at build time.
Goal
Provide a single module that exports all page metadata, enabling:
Type-safe access to page metadata
Easy iteration over all pages (for sitemap, feeds, navigation)
Consistent metadata structure across the app
Server-side only execution (no glob in the client bundle)
Relevant Docs
Remote functions - overview of query, form, command, and prerender
$app/server - API reference for query, prerender, etc.
Approach
Use a prerender remote function since page metadata only\n changes at build time. The result gets cached and prerendered during the\n build, so no runtime server is needed for the static site.
"},{"id":"/todos/update-readme-and-claude-md/","url":"/todos/update-readme-and-claude-md/","relPath":"/todos/update-readme-and-claude-md/","title":"Update README and CLAUDE.md","summary":"","image":"","tags":["Content"],"order":{"default":0},"authors":["ramil"],"wordCount":96,"date_published":"2026-01-26T00:00:00-00:00","content_html":"
Update README and CLAUDE.md
Ramil Karimovdeveloper
Update project documentation to reflect the current state after removing\n magazine/Timestripe content.
About Sveleton
Sveleton is a bootstrap project based on the Intuition Design System, built\n on the foundations of intuition.team and timestripe.com for quick-starting\n new media projects with everything ready for content writing, SEO, and more.
README.md
DONE - Updated with project description and commands.
CLAUDE.md
DONE - Updated with project overview, current structure, and todo page\n instructions.
"},{"id":"/todos/cleanup-all-warnings/","url":"/todos/cleanup-all-warnings/","relPath":"/todos/cleanup-all-warnings/","title":"Cleanup all warnings","summary":"","image":"","tags":["Cleanup"],"order":{"default":0},"authors":["ramil"],"wordCount":204,"date_published":"2026-01-23T00:00:00-00:00","updated":"2026-01-26T00:00:00-00:00","content_html":"
Cleanup all warnings
Ramil Karimovdeveloper
· Updated on
Fix all warnings from build process, svelte-check, and linting to achieve\n clean output.
Current Status
pnpm check: 0 errors, 0 warnings
pnpm build: 6 warnings (see below)
Build Warnings
1. tsconfig.json base config (2 warnings)
Cannot find base config file \"./.svelte-kit/tsconfig.json\"
Cause: esbuild runs in parallel with vite, starts before SvelteKit generates\n tsconfig.
Status: Added svelte-kit sync to build script but warning persists due\n to parallel execution. Harmless - build succeeds.
2. CSS :has() selector (4 warnings)
Expected identifier but found whitespace\nUnexpected \"has(\"
Location:src/styles/ids/settings.pcss:82
html:has(.theme_dark) { ... }
Cause: esbuild's CSS parser doesn't fully support :has() pseudo-selector.
Fix options:
Ignore (CSS works fine in browser, just parser warning)
Refactor to use JS-based theme switching instead of :has()
Use @supports or fallback approach
Recommended Approach
tsconfig warning: Leave as-is, it's harmless and resolves\n on build
:has() warning: Keep for now - modern CSS feature that\n works in browsers. Consider refactoring only if it causes actual issues.
"},{"id":"/todos/restore-meta-context/","url":"/todos/restore-meta-context/","relPath":"/todos/restore-meta-context/","title":"Restore meta context","summary":"","image":"","tags":["Features"],"order":{"default":-1},"authors":["ramil"],"wordCount":58,"date_published":"2026-01-23T00:00:00-00:00","content_html":"
Restore meta context
Ramil Karimovdeveloper
The getContext('meta') call in src/lib/meta.svelte was commented out to investigate build freeze.
Location
File: src/lib/meta.svelte:68
Task
Uncomment the getContext('meta') line
Verify build still works after restoring
If build freezes again, investigate why this call causes issues
"},{"id":"/todos/set-up-gh-actions/","url":"/todos/set-up-gh-actions/","relPath":"/todos/set-up-gh-actions/","title":"Set up GitHub Actions","summary":"","image":"","tags":["Infrastructure"],"order":{"default":0},"authors":["ramil"],"featured":true,"wordCount":128,"date_published":"2026-01-23T00:00:00-00:00","updated":"2026-04-17T00:00:00-00:00","content_html":"
Set up GitHub Actions
Ramil Karimovdeveloper
· Updated on
Configure GitHub Actions for automated builds, linting, and deployment.
![\"\"](\"/_app/immutable/assets/ramil.DJHUn9oQ.avif)
![\"\"](\"/_app/immutable/assets/1.B08UqseV.avif\")
![\"\"](\"/_app/immutable/assets/2.DfhhDRpk.avif\")
![\"\"](\"/_app/immutable/assets/3.DYEpEso-.avif\")
![\"\"](\"/_app/immutable/assets/4.JpEcJ6s5.avif\")
![\"\"](\"/_app/immutable/assets/5.CNHj6f07.avif\")
![\"\"](\"/_app/immutable/assets/6.BG0xQ_M_.avif\")
![\"\"](\"/_app/immutable/assets/7.qzVr7XA6.avif\")
![\"\"](\"/_app/immutable/assets/8.CZEqRP4h.avif\")
![\"\"](\"/_app/immutable/assets/9.BWXSlW2R.avif\")
![\"\"](\"/_app/immutable/assets/10.Dwgit1NQ.avif\")
![\"\"](\"/_app/immutable/assets/11.C-g1E7F8.avif\")
![\"\"](\"/_app/immutable/assets/12.Cqd34zKj.avif\")
![\"\"](\"/_app/immutable/assets/13.Ci0unIIK.avif\")
![\"\"](\"/_app/immutable/assets/14.s1pr6V8C.avif\")
![\"\"](\"/_app/immutable/assets/15.CEG6-L4i.avif\")
![\"\"](\"/_app/immutable/assets/16.CDkjZZGr.avif\")
![\"\"](\"/_app/immutable/assets/17.DvDwqxOj.avif\")
![\"\"](\"/_app/immutable/assets/18.Dee7U19m.avif\")
![\"\"](\"/_app/immutable/assets/19.WzwffIOX.avif\")
![\"\"](\"/_app/immutable/assets/20.BwsJAAyU.avif\")
![\"\"](\"/_app/immutable/assets/21.CH_QsB7h.avif\")
![\"\"](\"/_app/immutable/assets/22.D4fCJlWr.avif\")
![\"\"](\"/_app/immutable/assets/23.Bzro0e7N.avif\")
![\"\"](\"/_app/immutable/assets/24.BO1nLo_D.avif\")
![\"\"](\"/_app/immutable/assets/25.CuUcZXUN.avif\")
![\"\"](\"/_app/immutable/assets/26.QxgluLLu.avif\")
![\"\"](\"/_app/immutable/assets/27.CUfkOC7K.avif\")
![\"\"](\"/_app/immutable/assets/28.rhEwm6jV.avif\")
![\"\"](\"/_app/immutable/assets/29.DxL_SXvv.avif\")
![\"\"](\"/_app/immutable/assets/30.BrwxZ0TO.avif\")
![\"\"](\"/_app/immutable/assets/cover.DT5_Pqg3.avif)