SEO & Technical Health Agent
Broken links, slipping Core Web Vitals, and schema drift get caught and reported before they cost you rankings.
What it watches
Core Web Vitals, broken internal/external links, and structured-data validity across the site
What it does
Runs scheduled crawls and reports regressions, broken links, and schema errors with the affected pages listed
When it asks a human first
If a regression looks severe enough to affect indexing (a broken canonical, a site-wide 5xx, a robots.txt misconfiguration), it alerts you immediately outside the normal weekly report instead of waiting for the digest. That alert names the specific page or template where the problem was first detected, not just the category of problem, so fixing it doesn't start with a second investigation to find where it actually is.
How it works
- 1Crawls on schedule
Weekly by default, or after every deploy if wired into your CI/CD pipeline, with an initial full-site crawl at setup establishing the baseline every later run gets compared against.
- 2Tracks Core Web Vitals over time
LCP, INP, and CLS on key pages tracked as a trend, catching a slow regression, not just one bad reading — measured separately for mobile and desktop, since a regression showing up on only one is invisible if the two get averaged together.
- 3Finds broken links
Internal and external links checked site-wide, with affected pages listed by name. A redirect chain longer than a couple of hops gets flagged too, not just a hard 404.
- 4Validates structured data
Schema errors reported per page, so a rich-result eligibility loss does not go unnoticed — covering whichever schema types actually appear on the site (Product, Article, FAQ, LocalBusiness), not one generic check that misses type-specific required fields.
- 5Escalates the severe ones
A broken canonical, a site-wide 5xx, or a robots.txt misconfiguration alerts immediately, outside the weekly report — delivered directly (email or Slack, whichever you set at setup) rather than waiting to be noticed in a dashboard nobody happened to open that day.
What triggers it
- · Weekly scheduled crawl (or after every deploy if CI/CD-connected)
- · Core Web Vitals regress past your threshold
- · An internal or external link returns a broken status
- · Structured data fails schema validation
What you get
- · A weekly Core Web Vitals trend report
- · A list of broken links with the affected pages named
- · A schema-validation error log
- · An immediate alert for indexing-critical regressions
Not the right fit if
Not for sites intentionally blocked from indexing (private portals, staging environments) — there is no ranking signal to protect. It's also not worth setting up for a site that almost never changes — no new pages, no deploys, nothing added in months — since a regression tracker has nothing new to compare against and would return the same report week after week.
The math
At ~200 page crawleds/mo × 2 min, that is roughly 6.7 h/mo — an estimate, not a guarantee.
Frequently asked
It reports and, for a defined subset (like broken internal links), can propose a fix for your approval — same 'propose, don't silently change visible content' rule as the Content Ops Agent.
Ready to put seo & technical health agent to work?
Tell us about your setup and we'll confirm it fits within 48 hours.