How these labels are made

Every LHI asset tag is a small QR code printed on a physical object. Scanning it opens the page you came from. This is how a new one gets made.

The short version

Ask Claude Code to tag the object. It assigns the next free number, writes the description, rebuilds this site, deploys it, and hands back the exact string to encode. Then print the label and scan it before it goes on the object.

The rules that matter

Numbers are never reused. If an object is destroyed, its page becomes a tombstone

saying so. A recycled number means a label in a drawer silently points at the wrong thing, which is the failure this whole system exists to prevent.

Encode uppercase, and drop the https://. Uppercase puts the QR into alphanumeric

mode instead of byte mode, so it needs fewer modules and the printed code stays readable at the size of a microSD card. Phone cameras open a bare domain fine.

Scan the printed label with a real phone before it goes on the object. Not a

browser, not curl - a camera. This step gets skipped and it is the one that catches the problems.

Nothing sensitive on a public tag. Paths are guessable: one number follows the

last. No credentials, no customer names, no site addresses, no network detail, no pricing. Anything failing that test gets marked private and goes behind a login.

Why it is built this way

An earlier LHI QR system put 35 codes on printed product flyers pointing at a WordPress site. The site was compromised and returned an error on every URL for weeks. The hosting dashboard later showed 1,365 visitors during that window - real people scanning a flyer and getting a broken page, with nobody at LHI any the wiser.

So: these pages are plain static files with no software behind them to break or be broken into, and a scheduled job checks every tag URL on a timer and reports what it finds - whether or not anything is wrong. A QR code on a physical object is fire-and-forget. The only defence against silent failure is something that speaks up without being asked.

Where the real content lives

The source of truth is one markdown file per tag in the AI-Codex, not on this site. This site is generated from those files and can be thrown away and rebuilt at any time.