Quality — Structured, Fresh and Self-Demonstrating

The agent-readiness dimension that makes the rest verifiable — structured JSON-LD, freshness, retrievable markdown twins and self-demonstration.

What does the quality dimension mean?

Quality is the agent-readiness dimension that makes the other five verifiable: the evidence that what a site declares is structured, dated, cheap to retrieve and actually implemented. Agents Welcome added it to the field's five-category model so that a readiness claim is a fact an agent or the Audit can confirm, not a checkbox. Quality covers how structured your data is, how fresh it is, how cheaply it can be retrieved, and whether the site practices what it documents.

Which signals and standards does quality cover?

Four signals, each with a check an agent can run:

• JSON-LD / schema.org — JSON-LD 1.1 is a W3C Recommendation (W3C, 2020) and the format Google recommends (Google, 2025), and Google's policy permits markup only for content visible on the page (Google, 2026). schema.org's DefinedTerm — a word, name, acronym or phrase with a formal definition — ties each term to its definition and term set (Schema.org, 2026), the pattern this site's Lexicon uses for its 78 terms.
• Freshness — an explicit last-verified date on every record (last_verified; last_verified_date in the Crawler Registry): all 198 records in this site's Almanac carry one, dated between 15 June and 30 July 2026, and /api/updates plus the Atom feed /updates.xml publish what changed.
• Retrievable structure / markdown twins — clean .md twins, Vary: Accept and well-formed HTML keep the cost of retrieval near zero, so an agent parses without effort.
• Self-demonstration — the site implements every technique it documents, so its live artifacts are the proof: its self-audit scores 100/100 across 18 checks, and the test suite asserts that result on every run.

How does the Agent-Readiness Audit score quality?

Indirectly. The Audit has no quality category: json-ld (7 points) is scored under Content and passes when the page's first JSON-LD block parses, and no check scores freshness or self-demonstration yet (Agents Welcome, 2026). Quality is verified by refetching instead: the @graph validates, the last-verified dates are real and the markdown twins return, or they do not. This site ships an @graph on every page — WebPage, TechArticle and BreadcrumbList, plus a FAQPage node wherever a page shows a visible FAQ — 198 dated records across five datasets, and a self-audit of 100/100 that its tests lock in.

Quality — frequently asked questions

What makes structured data high quality rather than merely present?

Three properties: it parses as a valid @graph, it matches the visible page, since Google's policy forbids marking up content readers cannot see, and it carries a date an agent can weigh.

How fresh does agent-ready data have to be?

Fresh means dated and checked, not merely recent. Every Almanac record on this site carries a last-verified date, 15 June to 30 July 2026, and changes are published at /api/updates. Changing a date without changing the content is timestamp gaming, not freshness.

Can I verify this site's quality claims myself?

Yes. Fetch any page with Accept: text/markdown, read the @graph in its head, call /api/state-of-the-agentic-web for the dated records, or POST this site's URL to /api/audit: the score is 100 across 18 checks.

Related: schema.org for agents · implement JSON-LD · the Agentic Web Lexicon · adoption measured over time · audit your site

← Agent-Readiness · .md