Manifest position 415 · CrUX rank bucket 1000
https://www.merkur.de
Coverage complete
Coverage-complete agentic audit across homepage, section, article, search, and failure-state archetypes.
All 58 atomic check outcomes
| Principle / check | Status | Confidence | Evidence or reason |
|---|---|---|---|
respect-user-preferencesrespects-color-scheme | issues | high | Dark-mode screenshot remained a white consent surface and the CSS probe found no prefers-color-scheme rules or color-scheme declaration. F01 high: No user-preference theming |
respect-user-preferencesrespects-reduced-motion | issues | high | With prefers-reduced-motion: reduce matched, getAnimations() still reported a 500 ms fadeIn animation; CSS inspection found no reduced-motion media query. F01 high: No user-preference theming |
respect-user-preferencesrespects-contrast | pass | high | Forced-colors capture kept text, links, and the primary action visible with system-enforced contrast. |
implement-natural-interactionsview-transitions | issues | high | CSS/source probe found zero view-transition rules across the MPA navigation surface, so route changes remain abrupt. F02 low: Abrupt multi-page navigation |
implement-natural-interactionsscroll-driven-animations | pass | high | CSS probe found no animation-timeline rules and visual captures showed no scroll-linked effect, so the page does not implement scroll motion with main-thread scroll effects. |
implement-natural-interactionsphysical-gestures | not-applicable | high | No gesture-driven, swipe, carousel, or pull-to-reveal control was present on the representative news, section, article, or search templates. |
provide-guided-navigationscroll-state-aware-chrome | issues | high | The long news page uses static chrome; CSS probe found no scroll-state query, scroll timeline, or scroll-snap feedback implementation. F03 low: Long pages provide little position-aware guidance |
provide-guided-navigationanchored-positioning | not-applicable | high | No tooltip or edge-positioned popover was exposed in the representative paths. |
provide-guided-navigationdirects-attention | issues | high | Navigation across the multi-page templates has no directional/view transition or marker/highlight cue; CSS probe found zero view-transition and scroll-marker-related modern signals. F02 low: Abrupt multi-page navigation |
maximize-content-reduce-noiseno-intrusive-interruptions | issues | high | Every first-load screenshot is dominated by a full-viewport consent/subscription wall before article or search content can be read. F04 critical: Consent/subscription wall obscures all content |
maximize-content-reduce-noisesemantic-dismissible-primitives | issues | high | The visible consent modal is implemented inside a custom shadow-root .cmp_consentDialog using div-based .cmp_button controls, not a native dialog/button surface with declarative dismissal. F05 high: Consent controls use custom div semantics |
maximize-content-reduce-noisereduced-chrome | issues | high | The homepage profile counted 26 ad/iframe-like surfaces, while the HAR loaded 59 third-party requests and 2.73 MB of media before content interaction. F06 medium: Advertising and third-party chrome dominate the load |
adapt-to-the-form-factorresponsive-no-horizontal-scroll | pass | high | At 360x800 all four templates reported scrollWidth equal to clientWidth (360 px), viewport metadata present, and zero horizontal overflow. |
adapt-to-the-form-factorcomponent-level-responsiveness | issues | high | Rendered CSS probe found zero @container rules, so reusable news cards rely only on global layout behavior rather than container-level adaptation. F07 low: No component container queries |
adapt-to-the-form-factorinput-modality-aware | issues | high | Focus was visible on the first link, but 140 of 296 rendered controls/links were below 44 px in at least one dimension, including 34 px header controls and 39 px search buttons. F08 medium: Many interactive targets are undersized |
support-core-task-successclear-purpose-and-primary-action | pass | high | The homepage, section, article, and search each expose one descriptive H1 and template-specific titles; crawler captures preserve the news content hierarchy. |
support-core-task-successprimary-flow-completion | issues | high | Direct URLs return content and search results, but first-time readers must choose tracking or subscription in a full-screen wall before the primary reading/search journey is available. F09 high: First-time reading and search are blocked F04 critical: Consent/subscription wall obscures all content |
support-core-task-successclear-system-state-and-recovery | pass | high | A synthetic missing URL produced a clear “Inhalt nicht gefunden” state with address guidance, search, and service links for recovery. |
be-fast-and-stablegood-core-web-vitals | issues | high | Mobile Lighthouse measured LCP 3.94 s (outside the good <=2.5 s range), despite low CLS 0.005 and TBT 16.5 ms in that run. F10 high: Mobile LCP is outside the good range |
be-fast-and-stablevisual-stability | issues | high | Mobile layout observers measured CLS 0.2455 on search and 0.1283 on the politics section; the former exceeds the poor threshold and the latter exceeds good. F11 high: Search and section layouts shift materially |
be-fast-and-stableefficient-main-thread | pass | high | The trace found one 106 ms long task and 56 ms total blocking time over 11.3 s, while Lighthouse TBT was 16.5 ms; scripting did not dominate the measured load. |
be-fast-and-stableefficient-resource-delivery | issues | high | HAR recorded 90 requests and 3.99 MB; Lighthouse estimated 1,010 ms savings from render-blocking requests, and the HAR confirmed three VeryHigh parser-inserted stylesheets. F12 medium: Heavy render path and resource delivery |
be-fast-and-stabletrim-unused-and-duplicate-code | issues | high | Lighthouse estimated 132 KiB unused JavaScript, 29 KiB unused CSS, and 8 KiB avoidable legacy JavaScript; HAR recorded 25 scripts. F13 medium: Unused and legacy code ships on initial load |
be-inclusivenames-roles-labels | issues | high | Lighthouse failed link-name; direct probes found 83 empty links on the section, 89 on the article, and five homepage images without alt according to the image audit. F14 high: Numerous links have no accessible name F25 medium: Many crawlable links lack descriptive text |
be-inclusivesufficient-contrast | issues | high | Lighthouse accessibility audit failed color-contrast on the rendered page. F15 high: Text contrast failures |
be-inclusivestructure-and-focus | issues | high | Lighthouse failed heading-order; the section jumps from H1 directly to repeated H3 elements, including empty H3s. Focus itself was visible with a 2 px outline. F16 medium: Heading hierarchy skips levels |
be-inclusivelegible-text | pass | high | Rendered/crawler screenshots show readable German headlines and body text without clipping at desktop and narrow viewport; mobile layouts report no horizontal overflow. |
be-inclusivezoom-reflow-targets-and-media | issues | high | Viewport permits scaling and 360 px reflow has no overflow, but 140 of 296 rendered targets are below 44 px in one dimension. F08 medium: Many interactive targets are undersized |
follow-best-practicesno-console-errors | issues | high | Lighthouse explicitly failed errors-in-console during page load. F17 medium: Console errors occur on load |
follow-best-practicessound-document-and-assets | issues | high | Doctype, UTF-8 charset, and viewport are valid, but the image audit found 14 of 17 homepage images without width/height and seven oversized images. F18 medium: Images lack dimensions and are oversized |
follow-best-practicesbrowser-platform-hygiene | pass | high | Lighthouse best-practices scored 0.96; no deprecation, paste-prevention, notification-on-load, or geolocation-on-load failure was surfaced, and source probe found no permission call signatures. |
be-discoverabletitle-and-description | pass | high | All four templates have descriptive titles and rendered meta descriptions; the homepage, section, and search descriptions are also present in raw HTML. |
be-discoverablecrawlable-and-mobile-friendly | issues | high | Viewport and robots are valid and anchors carry hrefs, but Lighthouse failed link-name and direct probes counted 75–89 links with no discernible text/name on representative templates. F25 medium: Many crawlable links lack descriptive text |
be-discoverablecanonical-and-indexing-signals | pass | high | Public templates return 200, expose matching canonical URLs and index/follow; robots.txt allows public content and declares https://www.merkur.de/news.xml as the sitemap. Search is intentionally noindex/nofollow. |
be-discoverablestructured-and-shareable-metadata | issues | high | The article exposes Open Graph metadata but its sole JSON-LD object is typed WebPage rather than NewsArticle, reducing article-specific entity detail for rich results. F19 medium: Article schema is generic |
be-private-and-securesecure-transport-and-headers | issues | high | The page is HTTPS, but headers inspection found no CSP, HSTS, nosniff, frame defense, Referrer-Policy, or Permissions-Policy; five of six cookies lacked Secure. F20 high: Security headers and cookie protections are missing |
be-private-and-securedata-minimisation-and-third-parties | issues | high | HAR recorded 59 third-party requests transferring 3.41 MB; tracker enumeration found 19 third-party origins before consent choice, including ad/recommendation vendors. F21 high: Large third-party footprint before consent |
be-private-and-securein-context-permissions-and-modern-auth | not-applicable | high | No browser permission request occurred on load, and authentication/account registration was not part of the selected public-news core journey. |
be-private-and-securedefensive-browser-policies | issues | high | Header evidence found no HSTS, clickjacking defense, CSP/frame-ancestors, Referrer-Policy, or Permissions-Policy. Several cookies use SameSite=None without Secure. F20 high: Security headers and cookie protections are missing |
be-resilientprogressive-enhancement | pass | high | Discoverability captures show 85–92% raw-HTML coverage on public news templates with title/H1 preserved and no empty JS shell. |
be-resilientresilient-runtime-behaviour | pass | high | The custom consent wall remains contained and scrollable at 360 px; representative templates rendered without horizontal clipping and navigation remains ordinary crawlable links. |
be-resilientoffline-and-installable | not-applicable | high | Merkur.de is a public news publication rather than an intrinsically app-like workflow; offline/installability was treated as contextual. |
be-resilientnetwork-and-http-failure-states | pass | high | A missing route returns a clear branded not-found message and offers search and navigation recovery rather than a blank shell. |
be-internationalisedlang-dir-and-logical-properties | issues | high | The document correctly declares lang=de, but rendered CSS inspection found zero logical-property declarations against 545 physical left/right declarations, making bidi/writing-mode adaptation brittle. F22 low: CSS relies entirely on physical directions |
be-internationalisedlocale-aware-data | pass | high | The selected German-localized templates consistently render German-language dates, labels, and content under html lang=de; no foreign-locale product/currency flow exists. |
be-internationalisedtime-zone-correctness | not-applicable | high | The representative reading/search templates do not accept, schedule, or coordinate user event times across zones. |
be-trustworthyno-dark-patterns | issues | high | The consent wall gives a dominant ACCEPT ALL action and a subscription alternative, but no equally prominent reject option is visible; the primary content is blocked until a choice. F23 high: Consent and subscription choice is asymmetric |
be-trustworthyhumane-error-handling | pass | high | The only primary input is a labelled GET search field; the tested populated query returned results, while the not-found state clearly explains recovery without blaming the user. |
be-trustworthytrustworthy-input-assistance | not-applicable | high | The primary form is a single free-text news search; address, payment, sign-in, and sign-up autofill assistance does not apply. |
be-trustworthysafe-commercial-and-account-flows | issues | high | The blocking wall asks users to “SUBSCRIBE NOW” without showing price or commitment details in the decision surface, while acceptance of tracking is the visually dominant alternative. F23 high: Consent and subscription choice is asymmetric |
be-sustainableoptimised-assets | issues | high | Homepage image audit found seven oversized images, 11 legacy-format images, 14 missing dimensions, and 11 below-fold images without lazy loading. F18 medium: Images lack dimensions and are oversized F24 high: Background media and third parties consume most bytes |
be-sustainableno-wasteful-work | issues | high | Before content interaction the HAR transferred 2.73 MB across nine media requests and 3.41 MB third-party bytes, indicating avoidable background fetch work. F24 high: Background media and third parties consume most bytes |
be-sustainablethird-party-and-media-budget | issues | high | Third parties account for 59 requests and 3.41 of 3.99 MB total transfer; autoplay/background streaming media alone contributes 2.73 MB before the reader opts in. F24 high: Background media and third parties consume most bytes |
be-agent-readystructured-agent-capabilities | not-applicable | high | No declared agent-facing capability or transactional tool surface was found; public article content is nevertheless server-rendered and machine-readable. |
be-agent-readyon-device-inference | not-applicable | high | No on-device summarization or language-model feature is part of the current publication experience. |
be-memory-efficientno-leak-under-repeated-interaction | pass | high | After ten bottom/top feed traversals, heap self-size increased from 21.51 MB to 21.74 MB (about 1.1%) and DOM nodes sampled 4092→4096, which is bounded for lazy content activation. |
be-memory-efficientbounded-footprint | pass | high | Baseline heap summary reports 21.5 MB self-size and the live DOM sample reports 4,092 elements, proportionate to a long news homepage with 90 network requests. |
be-memory-efficientno-detached-dom-or-unbounded-listeners | pass | high | Before/after heap summaries show only 0.8% node-count growth (372,972→375,993) and no Detached* constructor among the largest retained constructors; live DOM grew by four nodes. |
Provenance
Canonical report: results/atomic/reports/0415-www_merkur_de.json
Report SHA-256: 8bad8a007d9730c1bd9adc951f6fccb6314d7bb1a59067ba3686752e4aa6ae10
Local retained report: runs/2026-07-17T17-27-24-856Z/atomic-reports/www_merkur_de/2026-07-20T16-02-49-335Z/report.json
Local evidence root: runs/2026-07-17T17-27-24-856Z/atomic-reports/www_merkur_de/2026-07-20T16-02-49-335Z
Catalog SHA-256: sha256:78ccfdb2d483f4c57d9dafed80fd86c6265585a56457c8dcfddc254b80fb44d7
Raw screenshots, HARs, traces, heaps, and other browser artifacts are retained at the local evidence root and intentionally are not committed. Artifact paths in the canonical report are relative to that root.