REFERENCE · scope stated in document

Primary-source register

1,770 words. Reference document; its scope and limits are stated inside it. Source: Research/SOURCE-REGISTER.md. Reference modules, shelf 4 of 8; library release 2026.09.12-g40.

Research snapshot: 9 September 2026. These 52 primary sources were consulted for the stated scopes: 50 search, analytics, web-platform and agent sources, plus two sources for the proposed CI workflow. A source supports only its stated claim, not every recommendation in the library. Operational workflows and acceptance gates are proposed agency practices, not undisclosed ranking rules.

Older live documentation can lag a dated update. In particular S04/S05 supersede S03 on dedicated AI reporting; S01 supersedes S21 on FAQ eligibility. S48 is intentionally version-pinned, not represented as the latest MCP security contract. No source publication date is invented when unavailable.

S01 — Google Search documentation updates

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: 2026-09-08

FAQ retirement; 2026 policy, feature and documentation changes. Changelog dates distinguish documentation changes from feature launches.

Open primary source

S02 — AI optimization guide

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

No special AI file or schema requirement for Google Search; llms.txt has no special Search role.

Open primary source

S03 — AI features and your website

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: 2025-12-10

Search eligibility and snippet controls. Its older reporting paragraph is superseded by S04/S05 for dedicated AI reporting.

Open primary source

S04 — Introducing Search Generative AI performance reports

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: 2026-06-03

Announcement updated with worldwide rollout on 2026-08-31; dedicated reports coexist with overall performance data.

Open primary source

S05 — Generative AI performance report (Search)

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

AI Overviews/AI Mode impressions; page, country, device and date dimensions. Availability requires sufficient data. No dedicated clicks or prompt dimension asserted.

Open primary source

S06 — Introducing AI Performance in Bing Webmaster Tools

Publisher: Microsoft Bing · Observed: 2026-09-09 · Document/announcement date: 2026-02-10

Public-preview AI citation reporting.

Open primary source

S07 — New AI visibility insights: Intents, Topics, Citation Share and Compare

Publisher: Microsoft Bing · Observed: 2026-09-09 · Document/announcement date: 2026-06-16

Citation Share is relative to all cited sites for a grounding query; topic/intent classifications and period comparisons.

Open primary source

S08 — Overview of OpenAI crawlers

Publisher: OpenAI · Observed: 2026-09-09 · Document/announcement date: not assigned

Separate OAI-SearchBot, GPTBot, ChatGPT-User and OAI-AdsBot roles; published verification feeds and user-fetch caveat.

Open primary source

S09 — Publishers and Developers FAQ

Publisher: OpenAI · Observed: 2026-09-09 · Document/announcement date: not assigned

ChatGPT referral UTM convention; search-access guidance.

Open primary source

S10 — Does Anthropic crawl data from the web?

Publisher: Anthropic · Observed: 2026-09-09 · Document/announcement date: 2026-04-07

ClaudeBot training, Claude-SearchBot search and Claude-User retrieval; published robots behavior and IP-information link.

Open primary source

S11 — Perplexity crawlers

Publisher: Perplexity · Observed: 2026-09-09 · Document/announcement date: not assigned

PerplexityBot search and Perplexity-User user-directed fetch; IP feeds and robots caveat.

Open primary source

S12 — Google common crawlers

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: 2026-07-14

Google-Extended is a robots product token, not a separate HTTP user agent; Gemini training/grounding scope, not Google Search.

Open primary source

S13 — Verify requests from Google crawlers and fetchers

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Published IP ranges and reverse/forward DNS verification procedures.

Open primary source

S14 — JavaScript SEO basics

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Google can render JavaScript; rendering, HTTP, canonical and initial-noindex considerations.

Open primary source

S15 — How to specify a canonical URL

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Canonicalization methods are signals with different strengths, not a universal mandate or a guarantee.

Open primary source

S16 — Robots meta tag, data-nosnippet and X-Robots-Tag

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Page-level indexing and snippet controls; crawl access is needed to read directives.

Open primary source

S17 — How Google interprets robots.txt

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Robots group matching and crawl directives; robots is not authentication.

Open primary source

S18 — Spam policies for Google web search

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Doorway, scaled-content, hidden-content and reputation abuse; policies also cover manipulation of generative Search responses.

Open primary source

S19 — Creating helpful, reliable, people-first content

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

People-first evaluation and quality questions, not an engine-exposed numeric E-E-A-T score.

Open primary source

S20 — General structured data guidelines

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Visible-content accuracy, feature eligibility, validation and no guaranteed rich result.

Open primary source

S21 — Changes to HowTo and FAQ rich results

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: 2023-08-08

Updated 2023-09-14: HowTo retirement effective 2023-09-13. Old FAQ eligibility is superseded by 2026 retirement in S01.

Open primary source

S22 — Review snippet structured data

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Authentic visible reviews; no fake or undisclosed incentivized reviews; self-serving LocalBusiness/Organization review-star restriction.

Open primary source

S23 — Merchant listing structured data

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Product/Offer eligibility, sale-duration fields and truthful commercial data. Not a promise of AI citations.

Open primary source

S24 — Creating a React app

Publisher: React · Observed: 2026-09-09 · Document/announcement date: not assigned

Frameworks can support client rendering, prerendering and server rendering according to route needs.

Open primary source

S25 — generateMetadata

Publisher: Next.js / Vercel · Observed: 2026-09-09 · Document/announcement date: 2026-08-25

Metadata API, streaming behavior and framework-version-specific usage.

Open primary source

S26 — Metadata and OG images

Publisher: Next.js / Vercel · Observed: 2026-09-09 · Document/announcement date: 2026-08-25

App Router metadata and social image implementation.

Open primary source

S27 — htmlLimitedBots configuration

Publisher: Next.js / Vercel · Observed: 2026-09-09 · Document/announcement date: 2025-10-03

Option introduced in 15.2; custom regex replaces default list; all-bot blocking changes streaming behavior.

Open primary source

S28 — Web Vitals

Publisher: Google web.dev · Observed: 2026-09-09 · Document/announcement date: 2024-10-31

Current LCP, INP, CLS thresholds and 75th-percentile field methodology; lab is not field evidence.

Open primary source

S29 — GEO: Generative Engine Optimization

Publisher: Aggarwal et al. / arXiv · Observed: 2026-09-09 · Document/announcement date: not assigned

Benchmark and deployed-engine experiments; findings depend on setup, domain, metric and model. Not a 2026 Google production ranking specification.

Open primary source

S30 — IndexNow documentation

Publisher: IndexNow · Observed: 2026-09-09 · Document/announcement date: not assigned

Changed-URL submission, ownership verification, batches up to 10000; success acknowledges receipt, not indexing.

Open primary source

S31 — Indexing API quickstart

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Restricted to JobPosting or BroadcastEvent embedded in VideoObject, not a universal website indexing endpoint.

Open primary source

S32 — Guidelines for representing your business on Google

Publisher: Google Business Profile · Observed: 2026-09-09 · Document/announcement date: not assigned

Real business representation, address/service-area eligibility and location restrictions.

Open primary source

S33 — Manage service areas for service-area and hybrid businesses

Publisher: Google Business Profile · Observed: 2026-09-09 · Document/announcement date: not assigned

Service areas versus customer-facing addresses; avoid invented storefronts.

Open primary source

S34 — Influencing title links

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Descriptive titles, automated selection and display truncation; no fixed character-count requirement.

Open primary source

S35 — Control snippets in search results

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Meta-description selection and truncation; no fixed character-count requirement.

Open primary source

S36 — Default channel group

Publisher: Google Analytics · Observed: 2026-09-09 · Document/announcement date: not assigned

Current AI Assistant channel; Google AI Overviews/AI Mode remain Organic Search; event/session/first-user scopes differ.

Open primary source

S37 — Avoid sending personally identifiable information

Publisher: Google Analytics · Observed: 2026-09-09 · Document/announcement date: not assigned

PII exclusion from URLs, titles, user inputs and campaign/event fields. Legacy UA code on this page is not adopted.

Open primary source

S38 — Site moves and migrations

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

URL mapping, redirects, monitoring and move sequencing.

Open primary source

S39 — Managing faceted navigation crawling

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Control unnecessary filter URL crawling; decide which facets need search visibility.

Open primary source

S40 — Tell Google about localized versions

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Hreflang implementation alternatives and reciprocal alternate relationships.

Open primary source

S41 — Regional differences in Google Search

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Region-, query- and feature-specific eligibility; documented on 2026-09-08.

Open primary source

S42 — Update to the site reputation abuse policy

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: 2026-08-28

Changes beginning 2026-08-30 distinguish EEA treatment; not permission to pursue abuse.

Open primary source

S43 — Search Console platform properties and social/video guide

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: 2026-07-29

Global platform properties for supported social/video platforms; Search visibility differs from native-platform analytics.

Open primary source

S44 — Analyze social and video content performance

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Connected/verified platform property workflows and Search performance interpretation.

Open primary source

S45 — Preferred Sources guide for web publishers

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

User-driven publisher preferences and button implementation; not a universal ranking guarantee.

Open primary source

S46 — Googlebot crawling size clarification

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: 2026-03-31

Googlebot fetch-size guidance: 2 MB for supported files except PDF, with headers included; separate resources fetched separately.

Open primary source

S47 — Model Context Protocol specification 2026-07-28

Publisher: Model Context Protocol · Observed: 2026-09-09 · Document/announcement date: 2026-07-28

Versioned agent integration protocol; implementation must negotiate supported version and enforce security.

Open primary source

S48 — MCP security best practices, 2025-11-25

Publisher: Model Context Protocol · Observed: 2026-09-09 · Document/announcement date: 2025-11-25

Version-pinned threat guidance: authorization, untrusted servers, SSRF and token/session risks; not the latest protocol contract.

Open primary source

S49 — MCP tools, 2026-07-28

Publisher: Model Context Protocol · Observed: 2026-09-09 · Document/announcement date: 2026-07-28

Tool interface metadata and human control; older examples can need migration.

Open primary source

S50 — Video SEO best practices

Publisher: Google · Observed: 2026-09-09 · Document/announcement date: not assigned

Watch-page eligibility, stable media resources and differences between video and ordinary web indexing.

Open primary source

S51 — actions/checkout official README

Publisher: GitHub Actions · Observed: 2026-09-09 · Document/announcement date: not assigned

Current checkout@v7 example and read-only repository permissions; proposed workflow only, not executed on GitHub.

Open primary source

S52 — actions/setup-python official README

Publisher: GitHub Actions · Observed: 2026-09-09 · Document/announcement date: not assigned

Current setup-python@v7 usage and Python version selection; proposed workflow only, not executed on GitHub.

Open primary source

Back to the shelf in the room · Reference modules