Manifest position 165 · CrUX rank bucket 1000
https://www.theguardian.com
Coverage complete
Coverage-complete atomic audit across six representative paths; no source supplied and no fixes applied.
Attempts
1 / 3
Judged checks
58 / 58
Blocked
0
Not run
0
This report has no published overall score. Blocked and not-run checks are not passes. A coverage-complete report means every check has a judged outcome; it does not mean every check passed.
All 58 atomic check outcomes
| Principle / check | Status | Confidence | Evidence or reason |
|---|---|---|---|
respect-user-preferencesrespects-color-scheme | issues | high | The dark-mode screenshot is pixel-identical in composition and colours to the default capture; the DOM reports html color-scheme: light and body background rgb(255,255,255). F1 medium: The main Guardian site is fixed to a light color scheme. |
respect-user-preferencesrespects-reduced-motion | pass | high | evaluate under prefers-reduced-motion: reduce found zero active animations |
respect-user-preferencesrespects-contrast | pass | high | high-contrast screenshot kept primary text, controls and consent choices visible; Lighthouse reported no contrast failure |
implement-natural-interactionsview-transitions | issues | high | The reduced-motion/source probe found no view-transition rules, and navigation is conventional full-document replacement. F2 low: Page and route changes do not use View Transitions. |
implement-natural-interactionsscroll-driven-animations | not-applicable | high | No scroll-linked decorative motion was observed on the audited templates, so there is no scroll animation implementation to assess. |
implement-natural-interactionsphysical-gestures | pass | high | homepage story carousel exposes a labelled forward control and remained bounded after repeated activation |
provide-guided-navigationscroll-state-aware-chrome | issues | medium | The long article template has many sections, while DOM/CSS inspection found no scroll-state container behaviour and the captured header is static. F3 low: Long article chrome provides no visible scroll progress or scroll-state response. |
provide-guided-navigationanchored-positioning | not-applicable | high | No tooltip or edge-sensitive anchored popover was present in the representative states; the full-screen modal does not need anchor positioning. |
provide-guided-navigationdirects-attention | pass | high | active Lifestyle section is visibly underlined and skip links guide focus directly to content/navigation |
maximize-content-reduce-noiseno-intrusive-interruptions | issues | high | Desktop and mobile first-load screenshots show the article/front content dimmed and inaccessible behind a fixed aria-modal dialog covering the viewport. F4 critical: A full-viewport consent-or-pay dialog obscures all content on first visit. |
maximize-content-reduce-noisesemantic-dismissible-primitives | issues | high | Article DOM inspection found gallery overlays as DIV role=dialog, while the consent system is an iframe wrapped by a custom aria-modal DIV. F5 medium: Rich overlays use custom role=dialog divs rather than native dialog semantics. |
maximize-content-reduce-noisereduced-chrome | pass | high | outside the first-visit dialog, the article and section templates devote the main region to journalism with compact global chrome |
adapt-to-the-form-factorresponsive-no-horizontal-scroll | pass | high | layout at 360x800 reports scrollWidth 360, clientWidth 360 and 0px horizontal overflow on home and article |
adapt-to-the-form-factorcomponent-level-responsiveness | pass | high | reused story/header components visibly reflow between desktop and 360px captures without clipping, even though no @container rules were found |
adapt-to-the-form-factorinput-modality-aware | pass | high | focus probe shows 3px focus box-shadows/outlines on sampled links, buttons and inputs; mobile controls are visibly touch-sized |
support-core-task-successclear-purpose-and-primary-action | pass | high | title, branded navigation and article H1 make the publication and reading purpose explicit; sign-in has a single clear continuation |
support-core-task-successprimary-flow-completion | issues | high | The article mobile capture remains blocked by the consent-or-pay dialog; the only primary actions are “Accept all” and “Reject all and subscribe”. F6 high: The primary reading journey cannot continue without accepting personalised advertising or purchasing Ad-Lite. |
support-core-task-successclear-system-state-and-recovery | pass | medium | representative 404 clearly says the page cannot be served, explains mistyped/outdated URLs, and links to the homepage |
be-fast-and-stablegood-core-web-vitals | issues | high | Lighthouse measured FCP 5.9s, LCP 8.8s, TTI 9.3s and performance 0.60; the raw-CDP trace under a lighter desktop profile measured LCP 2.10s, showing condition sensitivity but confirming mobile risk. F7 high: Mobile synthetic load performance misses the good LCP range. |
be-fast-and-stablevisual-stability | pass | high | raw-CDP layouts observed CLS 0 on home and article at 360px; Lighthouse CLS was 0.009 |
be-fast-and-stableefficient-main-thread | pass | high | trace found 2 long tasks, 61.88ms TBT, and Lighthouse TBT was 50ms, indicating bounded main-thread blocking despite slow loading |
be-fast-and-stableefficient-resource-delivery | issues | high | Lighthouse reports 820ms redirect savings, 153KiB cache-lifetime savings, font-display issues and 149KiB image savings; HAR records 108 requests and 1.14MB transferred. F8 medium: Critical delivery includes avoidable latency, cache and font/image inefficiencies. |
be-fast-and-stabletrim-unused-and-duplicate-code | issues | high | Lighthouse estimates 187KiB unused JS, 24KiB duplicated JS and 28KiB legacy JS; HAR records 44 script requests and 412,622 transferred script bytes. F9 medium: The initial load ships substantial unused, duplicated and legacy JavaScript. |
be-inclusivenames-roles-labels | issues | high | Lighthouse fails button-name and aria-required-children; the direct probe counted one unnamed button. F10 high: Some controls and composite widgets have incomplete accessible semantics. |
be-inclusivesufficient-contrast | pass | high | Lighthouse accessibility run reported no color-contrast failure and high-contrast screenshot remained legible |
be-inclusivestructure-and-focus | issues | high | Lighthouse fails the list audit and agent accessibility-tree audit, although the direct focus probe shows a visible 3px blue focus treatment on sampled controls. F11 medium: List and accessibility-tree structure is malformed in parts of the homepage. |
be-inclusivelegible-text | pass | high | article uses a clear heading hierarchy and readable column typography; 360px capture shows no clipped body copy |
be-inclusivezoom-reflow-targets-and-media | pass | high | viewport permits user scaling, 360px layouts reflow without overflow, sampled primary controls are touch-sized, and the detected video exposes a text track |
follow-best-practicesno-console-errors | pass | high | Lighthouse best-practices score is 1.00 and reported no console-error audit failure |
follow-best-practicessound-document-and-assets | issues | high | The images primitive found 98 of 100 images without width/height attributes. Although observed CLS was low in this run, the markup leaves stability dependent on CSS and timing. F12 medium: Most homepage images omit intrinsic dimensions. |
follow-best-practicesbrowser-platform-hygiene | pass | high | Lighthouse best-practices score is 1.00; no permission prompt appeared and Permissions-Policy disables sensitive capabilities |
be-discoverabletitle-and-description | pass | high | homepage and article DOM contain descriptive unique titles and meta descriptions |
be-discoverablecrawlable-and-mobile-friendly | pass | high | SSR anchors have real href values, viewport meta is present, and raw robots allows ordinary public content paths |
be-discoverablecanonical-and-indexing-signals | issues | high | Lighthouse fails canonical and robots.txt validity; the requested /sitemap.xml returns 404 while robots points to HTTP news/video sitemap URLs. The rendered final URL is /uk but canonical is the root. F13 medium: Homepage canonical and robots/sitemap signals are inconsistent for automated audits. |
be-discoverablestructured-and-shareable-metadata | pass | high | article includes NewsArticle JSON-LD plus matching Open Graph title, description, image and canonical URL |
be-private-and-securesecure-transport-and-headers | issues | high | Headers include HSTS and nosniff but CSP allows unsafe-inline and unsafe-eval. Cookie inspection found two first-party cookies without Secure and all seven set SameSite=None. F14 high: CSP and cookie settings leave avoidable browser-enforced security gaps. |
be-private-and-securedata-minimisation-and-third-parties | issues | high | HAR records 105 cross-origin requests by origin classification and 968KB, including 12 Sourcepoint and 10 Ophan requests; aborted Ophan URLs include attention, performance, experiments and consent-view events. F15 medium: The first-visit page makes a broad analytics and consent-related request footprint before a choice. |
be-private-and-securein-context-permissions-and-modern-auth | pass | high | no permission prompt appeared on load; sensitive permissions are disabled by policy; sign-in offers Google, Apple and email/passcode rather than requiring a password |
be-private-and-securedefensive-browser-policies | pass | high | HSTS preload, nosniff, SAMEORIGIN, Referrer-Policy and a restrictive Permissions-Policy are all present |
be-resilientprogressive-enhancement | pass | high | discoverability comparison found 100% content coverage in raw HTML, no empty app mount and no JS shell |
be-resilientresilient-runtime-behaviour | pass | high | consent dialog fits both desktop and 360px mobile viewports and focusable controls retain visible focus treatment |
be-resilientoffline-and-installable | not-applicable | high | This public news publication is content-first rather than an installable app. It provides a manifest but an offline application experience is not required for the primary reading model. |
be-resilientnetwork-and-http-failure-states | pass | medium | representative 404 is a complete server-rendered recovery page with explanatory copy and a homepage action |
be-internationalisedlang-dir-and-logical-properties | pass | high | document declares html lang=en and content reading order is correct; article markup is semantic and translation-readable |
be-internationalisedlocale-aware-data | pass | high | UK edition and article render publication data in locale-appropriate English day-month-year form |
be-internationalisedtime-zone-correctness | pass | high | article publication context explicitly uses the UK edition/time-zone label rather than presenting an ambiguous numeric timestamp |
be-trustworthyno-dark-patterns | issues | high | The modal offers “Accept all” versus “Reject all and subscribe” with no equally prominent free rejection path, coupling privacy refusal to a £5/month purchase. F16 high: The consent choice uses asymmetric accept-or-pay framing. |
be-trustworthyhumane-error-handling | pass | high | sign-in and newsletter fields use native email types and labels; no premature error was shown before interaction |
be-trustworthytrustworthy-input-assistance | issues | high | Article form inspection found type=email and a visible label but autocomplete is empty; the account sign-in email field correctly uses autocomplete=email. F17 low: The embedded newsletter email field omits an autocomplete token. |
be-trustworthysafe-commercial-and-account-flows | pass | high | sign-in clearly states account purpose and terms, offers federated and email options, and keeps password sign-in as an explicit alternative |
be-sustainableoptimised-assets | issues | high | The image probe reports 89 of 100 images without srcset, 100 legacy-format URLs, 14 below-fold images not lazy-loaded and two oversized images. F18 medium: Image delivery does not consistently use responsive source selection or modern formats. |
be-sustainableno-wasteful-work | issues | high | HAR captures repeated Ophan attention, performance, experiment and component events while the consent wall is still the only usable surface. F19 medium: Background analytics and experiment traffic starts on the blocked first-visit state. |
be-sustainablethird-party-and-media-budget | issues | high | Of 1.14MB transferred, fonts account for 252KB and Sourcepoint consent resources 202KB; the single Notice script is 94.9KB. F20 medium: Consent and font resources consume a disproportionate share of the first-load budget. |
be-agent-readystructured-agent-capabilities | not-applicable | high | The robots policy explicitly prohibits listed AI/agent crawlers and the publication does not declare an agent-facing capability intent, so structured agent actions are outside the intended surface. |
be-agent-readyon-device-inference | not-applicable | high | No user task in the audited publication requires built-in on-device inference, and the site explicitly restricts AI use of its content. |
be-memory-efficientno-leak-under-repeated-interaction | pass | medium | after 10 representative carousel advances heap self-size grew about 1.6% (29.69MB to 30.15MB), not evidence of unbounded retained growth in this bounded test |
be-memory-efficientbounded-footprint | pass | high | heap baseline was 29.69MB self-size with 473,334 nodes, proportionate to a very long media-rich news front |
be-memory-efficientno-detached-dom-or-unbounded-listeners | pass | high | baseline/post summaries show only modest bounded total growth after 10 carousel advances and no Detached* constructor among top retained populations |
Provenance
Canonical report: results/atomic/reports/0165-www_theguardian_com.json
Report SHA-256: a3b6bc5a313d502bad1ec5c5562ca53baac0f2c5b741643a76c1dde829616806
Local retained report: runs/2026-07-17T17-27-24-856Z/atomic-reports/www_theguardian_com/2026-07-18T21-20-50-631Z/report.json
Local evidence root: runs/2026-07-17T17-27-24-856Z/atomic-reports/www_theguardian_com/2026-07-18T21-20-50-631Z
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.