Editorial policy

Who writes it

Posts are written by the tragent team: the people who build tragent’s checks, its agent simulator and its scanner. They carry a team byline because each one draws on the product’s engineering work, not one person’s opinion.

Where the facts come from

• Primary sources first. Protocol specifications, the documentation of the platforms and AI companies involved, and standards bodies. Every post lists its sources at the end.
• Our own measurements. Where a post makes a claim about stores in general, it comes from scans we ran, described in our research methodology. We publish aggregates only and never name a store.
• Secondary sources, labelled. When the best available figure comes from a vendor’s own study or a news report, the post says so, and we do not repeat a number we could not trace to its origin.

How we use AI tools

We use AI writing tools to help research, outline and draft posts. Nothing is published as it comes out of a tool: a person on the team checks every factual claim against its source, runs every command and code sample, and edits the text before it goes live. Diagrams and charts are drawn from the post’s own data; we do not publish AI-generated photographs.

Our interest

tragent sells testing for agentic commerce, so we have an interest in the subject. We say so where it matters, we link to free and competing tools when they are the better answer, and a post’s advice should work whether or not you use tragent.

Dates and updates

Agentic commerce changes month to month. Every post shows the day it was published and, when it has been revised, the day its facts were last re-checked. An update date means a person reviewed the post; it never moves on its own.

Corrections

If something is wrong, write to [email protected]. We fix factual errors promptly, change the post’s update date, and note substantive corrections at the end of the post.