Manifest position 461 · CrUX rank bucket 1000
https://pubmed.ncbi.nlm.nih.gov
Coverage complete
Coverage complete. A browser anti-automation CAPTCHA blocked rendered inspection of the article detail archetype; its successful 200 raw HTML was inspected instead, and all registered checks were judged using the remaining rendered paths plus raw article evidence.
Attempts
3 / 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-condition screenshot is visually identical to the normal capture, while the reduced-motion computed probe reports body background rgb(255,255,255) and color-scheme: normal. F001 medium: The site does not provide a user-preference-driven dark theme. |
respect-user-preferencesrespects-reduced-motion | pass | high | The search page reported zero active animations under prefers-reduced-motion: reduce, while the normal probe observed animations. |
respect-user-preferencesrespects-contrast | pass | high | The forced-colors capture preserves visible text, search controls, outlines, and links with system colors. |
implement-natural-interactionsview-transitions | issues | high | The search CSS/DOM probe found no view-transition declarations despite pagination, filters, save panels, and result-state changes. F002 low: Search state and panel changes do not use View Transitions. |
implement-natural-interactionsscroll-driven-animations | pass | high | No scroll-linked animation is present, so the page does not use main-thread scroll handlers for a visual effect. |
implement-natural-interactionsphysical-gestures | pass | high | Search, pagination, and builder controls use ordinary platform click, scroll, select, and text interactions without custom gesture handling. |
provide-guided-navigationscroll-state-aware-chrome | pass | high | The mobile result view keeps search, sort, pagination, and result context clear without an obstructive sticky header. |
provide-guided-navigationanchored-positioning | not-applicable | high | No tooltip or viewport-edge anchored overlay was required in the representative core flow; existing menus were judged under semantic overlays. |
provide-guided-navigationdirects-attention | pass | high | Clear result count, pagination, ordered result numbers, blue titles, and search controls guide attention through the result list. |
maximize-content-reduce-noiseno-intrusive-interruptions | pass | high | Home, search, and advanced pages load without an interstitial or content-obscuring consent wall. |
maximize-content-reduce-noisesemantic-dismissible-primitives | issues | high | The search probe found 0 dialog, 0 popover, and 0 details elements while the DOM exposes Save, Email, Send to, citation, RSS, and filter panels with custom trigger/cancel wiring. F003 medium: Numerous overlays and action panels use custom controls instead of native dialog or popover primitives. |
maximize-content-reduce-noisereduced-chrome | pass | high | The home hero centers the search task and the result view gives most vertical space to citations rather than decorative application chrome. |
adapt-to-the-form-factorresponsive-no-horizontal-scroll | pass | high | At 360x800, scrollWidth and clientWidth are both 360 with 0 horizontal overflow; the viewport meta is present. |
adapt-to-the-form-factorcomponent-level-responsiveness | issues | high | The live stylesheet probe found no @container rules. Mobile screenshots show a good viewport adaptation, but reusable result, search, and builder components have no container-query behavior. F004 low: Responsive layouts rely on viewport rules rather than component-level adaptation. |
adapt-to-the-form-factorinput-modality-aware | pass | high | Lighthouse accessibility scored 1.0; visible mobile controls are touch-sized and labelled, and desktop controls use native focusable elements. |
support-core-task-successclear-purpose-and-primary-action | pass | high | The home page states the biomedical citation scope and prominently presents Search; results retain the query and result count. |
support-core-task-successprimary-flow-completion | pass | high | A GET search for cancer completed directly from the home search pattern to 5,667,691 results with pagination and citation actions. |
support-core-task-successclear-system-state-and-recovery | pass | high | The dedicated 404 page clearly says page not found and offers NCBI Home, Help Desk, and Search recovery choices. |
be-fast-and-stablegood-core-web-vitals | issues | high | Lighthouse measured LCP 3.0 s and FCP 2.3 s. A separate trace was faster at LCP 1.54 s, so the slower mobile condition is the actionable case; CLS remained 0. F005 medium: Throttled mobile LCP is above the good threshold. |
be-fast-and-stablevisual-stability | pass | high | Both the layout observer and Lighthouse measured CLS 0 with no recorded shifts. |
be-fast-and-stableefficient-main-thread | pass | high | Trace recorded 10.7 ms total blocking time with only two ~50 ms tasks; Lighthouse TBT was 90 ms, both within a responsive range. |
be-fast-and-stableefficient-resource-delivery | issues | high | HAR identified five parser-inserted VeryHigh-priority stylesheets; the DOM confirmed several parser scripts lack async/defer, and Lighthouse estimated 990 ms render-blocking savings. F006 medium: Five stylesheets and synchronous scripts create avoidable critical-path work. |
be-fast-and-stabletrim-unused-and-duplicate-code | issues | high | Lighthouse estimated 303 KiB unused JavaScript and 39 KiB legacy JavaScript; HAR measured 799,880 transferred script bytes. F007 medium: The result page ships substantial unused and legacy JavaScript. |
be-inclusivenames-roles-labels | pass | high | Lighthouse accessibility scored 1.0; home and advanced search inputs have explicit labels and icon-only controls expose accessible names. |
be-inclusivesufficient-contrast | pass | high | Lighthouse accessibility scored 1.0 and the forced-colors capture preserves readable system-color contrast. |
be-inclusivestructure-and-focus | pass | high | Lighthouse accessibility scored 1.0; pages expose h1/h2 hierarchy, main navigation, labelled controls, and a skip link. |
be-inclusivelegible-text | pass | high | Home, search results, and advanced builder use readable line lengths, spacing, and unclipped text at desktop and 360px. |
be-inclusivezoom-reflow-targets-and-media | pass | high | The viewport permits scaling, mobile content reflows with no overflow, and Lighthouse accessibility scored 1.0; no caption-requiring media was present. |
follow-best-practicesno-console-errors | pass | high | Lighthouse best-practices scored 1.0 and reported no console-error audit failure. |
follow-best-practicessound-document-and-assets | issues | high | Lighthouse failed the explicit width/height image audit, although the layout observer measured CLS 0 in this run. F008 low: Some images do not reserve explicit dimensions. |
follow-best-practicesbrowser-platform-hygiene | pass | high | Lighthouse best-practices scored 1.0; no permission prompt, paste blocking, or deprecation failure appeared in the representative flow. |
be-discoverabletitle-and-description | pass | high | Home, search, advanced, and article pages provide unique descriptive titles and meta descriptions. |
be-discoverablecrawlable-and-mobile-friendly | pass | high | Home raw-HTML coverage is 100%, links have real hrefs, viewport metadata is present, and robots.txt allows public content while excluding state-changing endpoints. |
be-discoverablecanonical-and-indexing-signals | pass | high | Home and article raw fetches returned 200; article and advanced pages provide canonicals. Search noindex is intentional for query combinations, while article is index,nofollow. |
be-discoverablestructured-and-shareable-metadata | issues | high | The raw 200 article HTML has descriptive title, description, canonical, and Open Graph tags, but contains zero application/ld+json blocks. F009 medium: Article pages provide social metadata but no structured scholarly-article data. |
be-private-and-securesecure-transport-and-headers | issues | high | The headers primitive found no Content-Security-Policy. Cookie evidence found _ga_DP2X732JSX with secure=false and a 400-day lifetime. HSTS, nosniff, DENY framing, and same-origin referrer policy were present. F010 high: Transport is strong, but CSP is absent and an analytics cookie is set without Secure. |
be-private-and-securedata-minimisation-and-third-parties | issues | high | HAR measured 43 third-party requests and 942,337 third-party bytes, including 446,606 bytes from Google Tag Manager and 106,887 from Qualtrics; trackers detected GTM and Google Analytics. F011 high: Analytics and survey dependencies dominate the transferred page weight. |
be-private-and-securein-context-permissions-and-modern-auth | pass | high | No geolocation/notification prompt fired on load. Authentication is external to the unauthenticated literature search and was not required for the core task. |
be-private-and-securedefensive-browser-policies | issues | high | Headers evidence found no Permissions-Policy and no CSP. HSTS and X-Frame-Options: DENY are correctly present. F012 medium: The response omits Permissions-Policy and CSP-based defenses. |
be-resilientprogressive-enhancement | pass | high | Home exposes 100% rendered-content word coverage in raw HTML and the article raw response contains title, abstract description, heading, and metadata without JavaScript. |
be-resilientresilient-runtime-behaviour | pass | high | Home search, results, advanced builder, mobile layout, and custom panels load without clipped core controls; a CAPTCHA prevented automated rendered article inspection but raw article content remained available. |
be-resilientoffline-and-installable | not-applicable | high | PubMed is an intrinsically online literature database rather than an installable offline app; an offline database was not an expected core capability. |
be-resilientnetwork-and-http-failure-states | pass | high | The 404 surface identifies failure and provides home, help, and search recovery links rather than a blank shell. |
be-internationalisedlang-dir-and-logical-properties | issues | high | The probe found html lang=en, but no margin-inline or padding-inline and no dir declaration; this makes future bidirectional/localized layouts more costly. F013 low: The globally used interface declares English but does not use CSS logical properties in the inspected search bundle. |
be-internationalisedlocale-aware-data | not-applicable | high | Representative public pages show bibliographic years and text but no locale-sensitive currency, number, or calendar formatting task. |
be-internationalisedtime-zone-correctness | not-applicable | high | No event scheduling, time-of-day, recurring interval, or time-zone-sensitive core data appeared on audited pages. |
be-trustworthyno-dark-patterns | pass | high | No consent wall, confirmshaming, upsell, disguised ad, forced continuity, or obstructed cancellation appeared in unauthenticated search and advanced flows. |
be-trustworthyhumane-error-handling | pass | high | Advanced fields are labelled and native required constraints are used in the saved-search panel; 404 recovery language is neutral and actionable. |
be-trustworthytrustworthy-input-assistance | pass | high | Search inputs use semantic search controls with labels; unauthenticated forms contain no address, payment, sign-in, or signup fields requiring specialized autocomplete tokens. |
be-trustworthysafe-commercial-and-account-flows | not-applicable | high | No checkout, subscription, payment, cancellation, or authenticated account-management flow was in unauthenticated scope. |
be-sustainableoptimised-assets | pass | high | Images transfer only 19,213 bytes and are primarily small SVG/data icons; the article page has no img elements requiring responsive raster delivery. |
be-sustainableno-wasteful-work | issues | high | The load made 43 third-party requests; GTM, federal analytics, Google Analytics, and Qualtrics scripts together account for most of the 978,202 transferred bytes. F014 medium: Background analytics and survey code adds substantial work unrelated to the immediate literature task. |
be-sustainablethird-party-and-media-budget | issues | high | Third parties transferred 942,337 of 978,202 bytes (96%); scripts alone transferred 799,880 bytes while the page contains primarily text citations. F015 high: Third-party JavaScript is disproportionate for a text-first search result page. |
be-agent-readystructured-agent-capabilities | issues | high | Discoverability found 100% raw-HTML coverage on the home page, but probes found no agent-tool registration or structured capability surface for search, filters, citation export, or save actions. F016 low: Machine-readable content is strong, but task capabilities are not exposed as structured agent tools. |
be-agent-readyon-device-inference | not-applicable | high | No page task clearly benefits from mandatory browser-side model inference; server search and static abstracts are appropriate. |
be-memory-efficientno-leak-under-repeated-interaction | pass | medium | After ten citation-trigger interactions, summarized heap was 21.58 MB versus a 20.30 MB fresh baseline (+6.3%); this bounded one-pass growth does not establish an unbounded leak. |
be-memory-efficientbounded-footprint | pass | medium | Fresh search heap self-size is 20.30 MB with 348,810 V8 nodes; Performance.memory used heap was 12.53 MB for a 3,249-element result page, proportionate to the feature set. |
be-memory-efficientno-detached-dom-or-unbounded-listeners | pass | medium | Neither heap summary surfaces a Detached* constructor among top retained constructors; jQuery wrapper count changed only 618 to 620 after interaction. |
Provenance
Canonical report: results/atomic/reports/0461-pubmed_ncbi_nlm_nih_gov.json
Report SHA-256: 8ec83aa04c01c8e2511dd4fbd7457807bb7194105c661ac6074108410776fec8
Local retained report: runs/2026-07-17T17-27-24-856Z/atomic-reports/pubmed_ncbi_nlm_nih_gov/2026-07-27T23-35-11-629Z/report.json
Local evidence root: runs/2026-07-17T17-27-24-856Z/atomic-reports/pubmed_ncbi_nlm_nih_gov/2026-07-27T23-35-11-629Z
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.