CtrlK
BlogDocsLog inGet started
Tessl Logo
PluginPublic

adobe/aem-eds-content-ops

Content operations skills for AEM Edge Delivery Services: page auditing, SEO optimization, AI search (GEO), WCAG accessibility, bulk metadata, structured data, sitemap validation, and content diffing. Complements the EDS development skills with operational tooling for content teams.

Quality

90%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Overview
Quality
Evals
Security
Files

performance-budget-rules.mdskills/performance-budget/references/

Performance budget: rules and a measurement without Chrome

Source for every rule here: Keeping it 100 and the AEM boilerplate, read in November 2026. Check them again before you rely on a detail.

The budget

  • The rule of thumb is that the aggregate payload before the LCP is displayed stays below 100 KB. On the bandwidth that PageSpeed Insights simulates for mobile, that usually gives an LCP under 1560 ms, which is what scores 100.
  • It is one budget for one host. A request to a second origin before LCP pays for a new connection as well.
  • Early hints, HTTP/2 push, preconnect and preload are part of the budget: they download things that the LCP doesn't need.
  • Lab results leave out redirects by default, but real visitors pay for each one. Report a redirect chain (/ to /en to /en/home) as a separate finding.

The three phases

PhaseWhat loadsCounts against the budget
EagerThe HTML, styles.css, aem.js, scripts.js, then the whole first section: its blocks' CSS and JS, and its first image, which is the LCP candidate.Yes
LazyStarts when the first section is shown: fonts.css, the other sections and their blocks, header and footer, lazy-styles.css, images as they scroll into view. Code the project controls, from the same origin.No
DelayedThird-party code: tag managers, analytics, consent, chat. The guideline is to start it at least three seconds after the LCP.No

The boilerplate takes the first <img> of the first section as the LCP candidate. When the real LCP element is not in the delivered HTML (it comes from a fragment, a JSON lookup or a service), the first block has to finish its changes to the DOM before the candidate is chosen. That lookup is then part of the budget.

For comparison: the boilerplate

Measured with Lighthouse 13.6.0 on https://main--aem-boilerplate--adobe.aem.live/ in November 2026, transfer bytes:

RequestBytes
LCP image (width=750, WebP)11,091
scripts/aem.js6,751
favicon.ico3,568
HTML2,725
scripts/scripts.js2,307
styles/styles.css1,917
CSS and JS of the two blocks in the first section2,335
Total30,694

A project starts with about 70 KB of headroom. What a page has on top of these numbers is what the project added.

Measuring with curl

Use this only when Lighthouse can't run. It gives sizes, not the LCP element and not the order of requests.

BASE="https://<branch>--<repo>--<owner>.aem.live"
for p in "<path>" /styles/styles.css /scripts/aem.js /scripts/scripts.js \
  /blocks/<block>/<block>.css /blocks/<block>/<block>.js "<lcp-image-path-and-query>"; do
  printf '%s\t%s\n' "$(curl -s --compressed -o /dev/null -w '%{size_download}' "$BASE$p")" "$p"
done

The LCP image comes from the user (DevTools or RUM). Take its URL with the query string from the <source> that matches a mobile viewport: the media pipeline sizes the image by width.

skills

performance-budget

references

performance-budget-rules.md

CHANGELOG.md

package.json

SKILL.md

tile.json