# Common Crawl: two-thirds of llms.txt files are templates

Published: 2026-08-31T22:03:52.907Z · Source: Common Crawl Blog (https://commoncrawl.org/blog/a-content-analysis-of-llms-txt-files-from-the-july-2026-crawl-archive)
Source date: 2026-08-31
Entities: llms-txt

Common Crawl analyzed 584,107 llms.txt files gathered in its July 2026 crawl archive and found 68% were produced by templates or SEO plugins — Wix alone accounts for 41% of the corpus, with All in One SEO and Yoast contributing smaller shares — while only about half fully follow the specification's structure of a title, summary line, and link sections. Automated screening flagged 3,793 files for steering language aimed at AI crawlers and 10 for "classic prompt injection" severity, but manual review confirmed only 4 files contained a genuine injection attempt.

Why it matters: The count suggests llms.txt adoption is running mostly on CMS-plugin autopilot rather than deliberate authoring — one more reason to doubt the file format is doing much to shape what AI crawlers see.

## What this answers

**How many llms.txt files did Common Crawl analyze?**

Common Crawl analyzed 584,107 llms.txt files gathered in its July 2026 crawl archive, finding roughly two-thirds were produced by templates or CMS plugins rather than written by hand.

**Do llms.txt files contain prompt injection attacks?**

Automated screening flagged thousands of files with promotional or steering language, but Common Crawl's manual review confirmed only 4 of the 584,107 files contained a genuine prompt injection attempt.


Canonical: https://anythingengineoptimization.com/item/2026-08-31-common-crawl-two-thirds-of-llms-txt-files-are-templates/
From Anything Engine Optimization (AEO Wire) — https://anythingengineoptimization.com/ · Standards: https://anythingengineoptimization.com/standards/