Protect (rule checks)

Deterministic SEO, schema, a11y, and recipe checks — rule-only, fast.

Inspector diff triage — side-by-side panels and classification matrix

Note

Protect is part of Creytix Inspector — QA and production readiness, detect and report, not auto-fix.

Overview

Protect is the fast, deterministic layer of Inspector: a set of rule checks for SEO metadata, structured schema, accessibility basics, and page "recipes" — the structural conventions a template is supposed to follow. Every check is rule-based — there's no model judgment call in the loop — so it runs quickly and gives you the same answer every time for the same input.

How it works

Run Protect against a page or a batch. It checks metadata, schema markup, accessibility attributes, and structural conventions against fixed rules.
Get a pass/fail report per rule, not a vague quality score.
Fix or override. Protect tells you what failed; your team decides whether to fix it or explicitly accept it.

Steps

•
Point Protect at a page or route
Run a check against a single page during editing, or a batch before a release.
•
Read the rule-by-rule report
Each check reports pass or fail against a specific, named rule — title length, missing alt text, schema shape, and so on.
•
Fix what needs fixing
Address failures in the content or template. Protect doesn't touch the page itself.
•
Re-run before you ship
Re-check after fixes, and let the CI gate hold the line on anything still failing.

Capabilities

SEO checks — title length, meta description presence, canonical tags, and related metadata rules.
Schema checks — structured data shape against what the page type expects.
Accessibility checks — the mechanical basics: alt text, contrast-adjacent attributes, focus and touch-target rules.
Recipe checks — a page's structure against the conventions its template is supposed to follow.

Limits & honest scope

Protect is rule-only by design — that's what makes it fast and repeatable.