Manifest position 585 · CrUX rank bucket 1000
https://www.amazon.fr
Coverage complete
Coverage-complete anonymous audit. Authenticated account management, payment submission, order placement and seller surfaces were excluded because they require credentials or create transactional risk.
All 58 atomic check outcomes
| Principle / check | Status | Confidence | Evidence or reason |
|---|---|---|---|
respect-user-preferencesrespects-color-scheme | issues | high | The dark-mode screenshot is visually identical to the light capture, and the CSS probe found no color-scheme declaration or prefers-color-scheme rule. F01 medium: The site does not honour the user’s dark colour-scheme preference. |
respect-user-preferencesrespects-reduced-motion | pass | high | Emulated prefers-reduced-motion: reduce; getAnimations() returned none and the home CSS contains a reduced-motion rule. |
respect-user-preferencesrespects-contrast | pass | high | Forced-colors capture retained visible text, controls, outlines and content with strong system contrast. |
implement-natural-interactionsview-transitions | issues | high | DOM/CSS probes on home, search and product found no view-transition rules; page changes are abrupt full navigations. F02 medium: Route and major state changes do not use View Transitions. |
implement-natural-interactionsscroll-driven-animations | pass | medium | No scroll-linked decorative motion was observed on sampled paths, and probes found no scroll-timeline anti-pattern requiring migration. |
implement-natural-interactionsphysical-gestures | pass | medium | Carousels and standard scrolling remained browser-driven; no sampled core gesture fought native scrolling or required custom pointer handling. |
provide-guided-navigationscroll-state-aware-chrome | issues | high | Desktop and mobile captures show two persistent navigation rows; CSS probes found no scroll-state query or scroll-timeline signal. F03 medium: Dense navigation chrome does not respond to scroll state. |
provide-guided-navigationanchored-positioning | issues | high | The home probe found 27 dialog-role elements and no anchor-name/position-anchor signal; the visible consent surface is a FORM with role=dialog. F04 low: Overlay infrastructure relies on many custom dialog-role containers rather than anchored platform primitives. |
provide-guided-navigationdirects-attention | pass | high | Search/product templates provide breadcrumbs, skip links and clearly labelled result/product destinations. |
maximize-content-reduce-noiseno-intrusive-interruptions | issues | high | At 1440x900 the consent dialog begins around y=437 and occupies the entire lower half; on the product page it covers product details and buying controls. F05 high: The first-load consent surface obscures most primary content. |
maximize-content-reduce-noisesemantic-dismissible-primitives | issues | high | The probe identifies #cos-banner as a FORM with role=dialog, not <dialog>; no aria-modal is present. F06 medium: The consent overlay is not implemented with the native dialog primitive. |
maximize-content-reduce-noisereduced-chrome | issues | high | Desktop captures show two dense navigation rows plus a horizontally clipped card rail; the home page has no single content heading and starts with competing promotional cards. F07 medium: Navigation and merchandising chrome crowd the content. |
adapt-to-the-form-factorresponsive-no-horizontal-scroll | issues | high | The mobile layout primitive reported no viewport meta, a 1000px CSS content width, 360px visual viewport and 640px horizontal overflow; the screenshot shows a clipped desktop canvas. F08 critical: The anonymous home page does not reflow to a 360px viewport. |
adapt-to-the-form-factorcomponent-level-responsiveness | issues | high | The home probe found a container-query signal, but search and product probes found none and both use desktop-sized structures without a viewport meta tag. F09 medium: Container-level responsiveness is inconsistent across templates. |
adapt-to-the-form-factorinput-modality-aware | issues | high | Lighthouse failed target-size and found 15px-high footer links; probes counted 13-22 sub-24px targets per template and many focused controls computed outline:none. F10 high: Many controls are too small or lose a visible focus indicator. |
support-core-task-successclear-purpose-and-primary-action | pass | high | Search clearly labels results and product detail exposes title, price, stock and purchase action; empty basket explains its state. |
support-core-task-successprimary-flow-completion | pass | medium | Anonymous search -> listing -> product -> basket/sign-in boundary was reachable without a dead end; order placement was excluded to avoid a transaction. |
support-core-task-successclear-system-state-and-recovery | pass | high | Empty basket gives sign-in/sign-up/deals options and the 404 page explains the failure with a home link. |
be-fast-and-stablegood-core-web-vitals | pass | medium | Trace measured LCP/FCP 1,924ms, one 77ms long task and 27ms TBT; mobile CLS observer measured 0. INP field data was unavailable. |
be-fast-and-stablevisual-stability | issues | high | The image audit found 39 of 40 images missing width/height. Observed CLS was 0 in this capture, but reserved space is absent for late or alternate loads. F11 medium: Images are shipped without intrinsic dimensions, creating avoidable shift risk. |
be-fast-and-stableefficient-main-thread | pass | high | Trace found one long task, longest 77.3ms, with 27.3ms total blocking time over 7.1s. |
be-fast-and-stableefficient-resource-delivery | issues | high | HAR: 131 requests and 2,388,945 transferred bytes, including 739,736 font bytes, 648,982 script bytes, six VeryHigh parser-inserted stylesheets and two parser scripts confirmed without async/defer. F12 medium: The critical path carries many blocking resources and a heavy font/script payload. |
be-fast-and-stabletrim-unused-and-duplicate-code | issues | high | HAR records 26 scripts (648,982 bytes) and 114 third-party-origin requests; a video-card bundle alone transfers 171,088 bytes. F13 medium: JavaScript and third-party delivery are disproportionate for the first viewport. |
be-inclusivenames-roles-labels | issues | high | Lighthouse found invalid aria-expanded on the searchbox, aria-setsize on an incompatible list, and a basket link whose visible “0” is missing from its accessible name. F14 high: Accessible names and ARIA roles are inconsistent. |
be-inclusivesufficient-contrast | pass | high | Lighthouse accessibility score was 0.93 without a colour-contrast failure; forced-colors capture remained legible. |
be-inclusivestructure-and-focus | issues | high | The home has no H1; the product template exposes multiple H1 elements. Probe focus checks found many active controls with outline:none, including navigation and submit controls. F15 high: Heading hierarchy and focus visibility are inconsistent. |
be-inclusivelegible-text | pass | high | Desktop and product screenshots show readable body and heading text with no visible clipping in the sampled desktop condition. |
be-inclusivezoom-reflow-targets-and-media | issues | high | There is no viewport meta tag, the 360px capture retains a 1000px layout, and Lighthouse failed touch-target sizing. F16 high: Mobile reflow and target sizing do not meet an inclusive baseline. |
follow-best-practicesno-console-errors | issues | high | Lighthouse captured a console.error for a late-loading @m/mash module from an Amazon JavaScript resource. F17 medium: The page logs a browser console error on load. |
follow-best-practicessound-document-and-assets | issues | high | Lighthouse found multiple footer images rendered at 1.49 aspect ratio from 2.69-ratio sources; the image audit also found 39 missing intrinsic dimensions. F18 medium: Image asset sizing is unsound. |
follow-best-practicesbrowser-platform-hygiene | issues | high | Lighthouse reports use of the deprecated Shared Storage API from the main document. F19 low: A deprecated browser API remains in use. |
be-discoverabletitle-and-description | issues | high | The search probe found a descriptive title but description:null. Home and product pages do include descriptions. F20 medium: Search result pages omit a meta description. |
be-discoverablecrawlable-and-mobile-friendly | issues | high | Discoverability measured 0% raw-to-rendered content coverage, isJsShell:true, no raw title/H1/description, and a blank crawler screenshot; Lighthouse also found javascript:void(0) links. F21 critical: Non-JavaScript crawlers receive an empty page and some links are not crawlable. |
be-discoverablecanonical-and-indexing-signals | issues | high | Home/product expose canonicals, but search has none, no hreflang links were found, and /sitemap.xml returned 404. F22 high: Indexing signals are incomplete across templates. |
be-discoverablestructured-and-shareable-metadata | issues | high | No JSON-LD was found on home or product; product has no Open Graph tags, while home only exposes og:description. F23 high: Product and home entities lack complete structured/share metadata. |
be-private-and-securesecure-transport-and-headers | issues | high | Headers primitive found no CSP, HSTS, nosniff, frame protection, Referrer-Policy or Permissions-Policy on the main document. Cookie audit found 8/8 flagged, including non-Secure values and session-id cookies without HttpOnly. F24 critical: The main document lacks baseline security headers and sets weak cookies. |
be-private-and-securedata-minimisation-and-third-parties | issues | high | The consent copy names 129 third parties; HAR records 114 third-party-origin requests and 1,889,978 third-party bytes, including ads and telemetry origins. F25 high: The initial page has a broad third-party and advertising footprint. |
be-private-and-securein-context-permissions-and-modern-auth | issues | high | Sign-in exposes username autocomplete but the password input has no current-password token; no passkey/WebAuthn affordance was visible in the anonymous flow. F26 medium: The observed authentication flow does not expose a modern passkey path and password autocomplete is incomplete. |
be-private-and-securedefensive-browser-policies | issues | high | The main response had no clickjacking protection, Referrer-Policy, Permissions-Policy, CSP/Trusted Types policy or HSTS according to the headers primitive. F27 high: Browser-enforced defensive policies are absent on the main document. |
be-resilientprogressive-enhancement | issues | high | Raw HTML was only 2,161 bytes with zero content tokens while the rendered page had 384; browser view was populated and crawler view blank. F28 critical: Core content is unavailable without JavaScript to a plain crawler. |
be-resilientresilient-runtime-behaviour | pass | medium | Sampled search, product, cart and 404 states rendered actionable content; overlays remained attached within the desktop viewport. |
be-resilientoffline-and-installable | not-applicable | high | Amazon retail depends on live price, inventory, identity and checkout state; offline/installability was judged out of scope for this public commerce audit. |
be-resilientnetwork-and-http-failure-states | pass | high | A real 404 returned a concise error explanation and a clear home recovery link rather than a blank shell. |
be-internationalisedlang-dir-and-logical-properties | issues | high | Pages set lang=en-gb, but no dir attribute or hreflang links are present and CSS probes found no logical-property signal on the sampled templates. F29 medium: Locale markup is only partially internationalisation-ready. |
be-internationalisedlocale-aware-data | pass | high | Observed prices use euros and locale formatting (€14.39 displayed as €14 with 39 cents); delivery copy uses a localized weekday/date. |
be-internationalisedtime-zone-correctness | not-applicable | medium | No user-scheduled events or time-zone-sensitive timestamps were present in the representative anonymous shopping journey. |
be-trustworthyno-dark-patterns | pass | medium | Consent offers adjacent Accept and Decline actions without confirmshaming; product price, delivery and stock are visible before basket action. Intrusiveness is reported separately. |
be-trustworthyhumane-error-handling | pass | medium | The sampled sign-in uses a single clearly described email/phone step; basket and 404 recovery states are plain and actionable. Destructive or payment validation was not triggered. |
be-trustworthytrustworthy-input-assistance | issues | high | The observed email field uses autocomplete=username, but the password field returns an empty autocomplete value. F30 medium: The sign-in password field omits the expected autocomplete token. |
be-trustworthysafe-commercial-and-account-flows | pass | medium | Product price, delivery, stock and quantity are visible; basket clearly states it is empty and sign-in is required before protected account work. Transaction submission was not attempted. |
be-sustainableoptimised-assets | issues | high | Image audit: 32 legacy-format images, zero modern-format images, 38 missing srcset, 24 below-fold images not lazy-loaded and 39 missing dimensions. F31 high: Image delivery misses modern responsive optimisation. |
be-sustainableno-wasteful-work | issues | high | HAR includes 131 requests, telemetry beacons and 4 media requests; a 171KB video-card bundle loads on the initial page despite no user interaction. F32 medium: First-load background and media work is not tightly budgeted. |
be-sustainablethird-party-and-media-budget | issues | high | Third-party origins transfer 1.89MB; fonts transfer 739,736 bytes, scripts 648,982 bytes, and media code/resources load before interaction. F33 high: Third-party, font and media costs dominate the page budget. |
be-agent-readystructured-agent-capabilities | issues | high | No WebMCP/modelContext signal was found; robots.txt disallows GPTBot, ClaudeBot, OAI-SearchBot, ChatGPT-User and other agents, while non-JS content coverage is 0%. F34 medium: Commerce capabilities are not exposed safely to agents, and agent crawlers are explicitly blocked. |
be-agent-readyon-device-inference | not-applicable | high | On-device generative inference is not necessary for the representative search, product and basket tasks; absence was not penalised. |
be-memory-efficientno-leak-under-repeated-interaction | pass | medium | After ten menu open/close attempts, heap self size changed from 31.61MB to 31.84MB (+0.23MB, 0.7%) and node count by +776, consistent with bounded variance rather than an unbounded leak. |
be-memory-efficientbounded-footprint | pass | medium | Baseline heap summary was 31.61MB across 461,370 nodes; substantial but proportionate to a large commerce homepage and stable after interaction. |
be-memory-efficientno-detached-dom-or-unbounded-listeners | pass | medium | Before/post heap summaries show no Detached* constructor among top retained constructors and only 0.7% self-size growth after repeated menu interaction. |
Provenance
Canonical report: results/atomic/reports/0585-www_amazon_fr.json
Report SHA-256: 3388f93b3532a1dbc2211190b4fccdc3aa9ca0930c116e41b4686b5a094b9a1b
Local retained report: runs/2026-07-17T17-27-24-856Z/atomic-reports/www_amazon_fr/2026-07-21T18-28-52-107Z/report.json
Local evidence root: runs/2026-07-17T17-27-24-856Z/atomic-reports/www_amazon_fr/2026-07-21T18-28-52-107Z
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.