Opens in a new tabSkip to content
Agent LighthouseAgent Lighthouse

    Searches the text of every published page. The evidence sources themselves are not in this index — search all of them on the trusted sources page.

    GitHub ↗
    Browse checks and page contents
    answer-readiness/text-fragment-addressability

    Text-fragment citation addressability

    What it checks

    Determines whether a citing surface can construct a working #:~:text= deep link to the page’s actual answer sentences. Hard-fails on the documented Document-Policy: force-load-at-top opt-out header. It then simulates the spec’s matching algorithm over the parsed DOM, to prove three things about each candidate answer span. The span is contained in a single block-level element. It is unambiguous, or disambiguable with a same-block prefix or suffix. And it is free of characters that break normalization. Outputs the working fragment URLs as a fix artifact.

    Why it matters

    Google Search auto-generates text-fragment URLs to land users on the exact featured-snippet text (S12), and the spec requires each of prefix/start/end/suffix to match within a single block-level element (S2, S3). When an answer sentence is fragmented across block boundaries, or the header opt-out is set, the fragment silently fails and the link degrades to page-top (S3). Falsifiable and directly testable: take the citing surface’s own generated URL, load it, and observe whether the browser scrolls and highlights. Two failure classes are binary and deterministic — the opt-out header, and a start string that straddles two blocks.

    Evidence

    • Text fragments — Google / web.dev (vendor-doc, URL verified 2026-08-20)
    • Confirms a shipped answer-surface consumer: “Clicking a featured snippet takes the user directly to the featured snippet text on the source web page. This works thanks to automatically created Text Fragments URLs.” Support: Chrome 89+, Edge 89+, Firefox 131+, Safari 18.2+. Restates the boundary rule: “Each of prefix-, start, end, and -suffix can only match text within a single block-level element, but full start,end ranges can span multiple blocks.” Opt-out header: Document-Policy: force-load-at-top.
    • Playwright: Auto-waiting / Actionability checks — Microsoft (vendor-doc, URL verified 2026-08-20)
    • Before click, check, fill and selectOption, Playwright enforces five checks. Visible: a non-empty bounding box, and not visibility:hidden. Stable: the same bounding box over 2 animation frames. Receives Events: the element is the hit target at the action point, so overlays cause failure. Enabled: not [disabled] or aria-disabled. Editable: not readonly or aria-readonly. Fill requires visible+enabled+editable. This is the exact gate every Playwright-based agent (Playwright-MCP, browser-use, most CUA harnesses) passes through, so each check is a directly testable site-side failure cause.
    • MCP Specification 2026-07-28 — Versioning and Compatibility — Model Context Protocol (spec, URL verified 2026-08-20)
    • ‘There is no negotiation handshake.’ Terminology: Modern = 2026-07-28+ (per-request _meta); Legacy = 2025-11-25 and earlier (initialize handshake). Unsupported version MUST return error code -32022 with data.supported[] and data.requested. Verbatim compatibility matrix: Modern client + Legacy server = fails. Legacy client + Modern server = fails. Only dual-era implementations bridge. Extensions negotiated via capabilities.extensions map with mandatory reverse-DNS prefix.
    • MCP Specification 2026-07-28 — Tools — Model Context Protocol (spec, URL verified 2026-08-20)
    • tools/list result set MUST NOT vary per-connection or as a side effect of other requests (MAY vary by authorization). Servers SHOULD return tools in deterministic order — rationale given verbatim: enables client caching and ‘improves LLM prompt cache hit rates’. inputSchema MUST be a valid JSON Schema object (not null); defaults to JSON Schema 2020-12. Tool names SHOULD be 1-128 chars, case-sensitive, only [A-Za-z0-9_.-], unique within a server. Full x-mcp-header constraint list including static-reachability rule (chain of only properties keys; never through items/oneOf/anyOf/allOf/not/if/then/else/$ref). Clients MUST exclude violating tools from tools/list. If outputSchema present, servers MUST conform. Clients MUST treat annotations as untrusted.

    How it scores

    Tier per evidence policy: scored — grade A meets the A/B bar required for scored audits.

    Example failure

    A CDN-managed site sends Document-Policy: force-load-at-top from a hardening template. Every AI and Search deep link into its documentation lands at the top of a 6,000-word page instead of the cited sentence. Second class. A spec page renders ‘Maximum payload size is’ in a <p>, and ‘10 MB’ in an adjacent <span class="value"> inside its own <div>. No single-block start string covers the sentence. No fragment can address the one fact the page exists to state.

    Sources