For automated readers

You may

• Read every public page here, and follow the links in our sitemap.
• Quote and cite what we publish, with attribution and a link back to the page you took it from.
• Summarise who WorkTrust is and what it does, in answers to people who asked.
• Check a single credential a person has shared with you, at the address they gave you.
• Use /llms.txt and /agents.txt as the short, current description of the product. They are generated from the site, so they do not go stale.

You may not

• Reproduce this site, its copy, its structure or its design as a product of your own, or as a template for one.
• Collect our content in bulk for a commercial dataset, a model's training corpus, or resale. Reading is not mining, and we reserve text-and-data-mining rights (see below).
• Train, fine-tune or evaluate a commercial model on this site's content without a written agreement with us.
• Enumerate, guess or sweep verification IDs. Each credential address is shared by its holder on purpose and can be withdrawn; /verify and /api/verify are rate-limited per caller, and a guess costs a hash rather than a lookup.
• Test our authentication, attempt sign-ins, replay sessions, or probe for weaknesses. If you are looking for a vulnerability, write to us instead. We would rather hear from you than from a log.
• Present yourself as WorkTrust, issue anything that looks like a WorkTrust credential, or imply a verification we did not make.

The same rules, where machines look

These four addresses say the same thing as this page, in the formats automated readers already parse. They are generated from the site itself, so they cannot drift from what is written here.

• /agents.txt — this policy and what the product is, as plain text
• /llms.txt — the same, at the address language models look for
• /robots.txt — what may be crawled, and the sitemap
• /.well-known/tdmrep.json — our text-and-data-mining reservation, machine-readable, under the W3C TDM Reservation Protocol