Two mechanisms are not in the guides.

The playbook, the harness guide and the workbook are complete on their own. They are also the part of our work the market already knows, which is why they are free. Two mechanisms sit behind them. This page names each one and what it detects. It does not publish how it is built, because that is the part you cannot copy from a document. Both are shown the same way: on your estate, in a working session, against a failure from your own delivery history rather than ours.

The two mechanisms.

Two gates.

P4

One gate proves the generator. A second gate takes the emitted product and starts it the way a person would. They are not the same gate run twice: the second exists because the first has the blind spots of whatever it runs.

What it detects: a green build that shipped something the runtime does not accept.

How we know it is real: one of our own pages answered 404 in production while every generator check was green, because the local server had been answering the missing page with a fallback. A build gate cannot see that. A product gate can.

A false pass is a defect class.

P5

A test that passes for the wrong reason is treated as a named defect, not bad luck. A green suite is not evidence, and a skip is not a pass. Every assertion has to be shown able to fail before it is counted.

What it detects: a suite that is green because it never ran the thing it claims to check.

How we know it is real: five of our own checks were found reporting false passes in a single review, and a sixth was found later. One skipped on every run because the path it looked for was the path it was asserting to be missing.

What the session covers.

One working session, on your systems, with your own recent delivery to hand. Four things happen in it, and you leave with the fourth whether or not we work together again.

1. We find one real false pass.

We look for a check in your estate that is green for the wrong reason. Most estates have at least one, and the reader usually recognises it before we finish the sentence.

2. We split your gates by layer.

Which of your checks read the artifact with the same assumptions as the thing that produced it, and which start the product the way a person does. The split is usually unequal.

3. We run the mechanism on one case.

Not a demonstration on our systems. The mechanism is applied to one of your failures, in front of you, and you watch it either find the defect or fail to.

4. You leave with the written rule.

The decision rule for the case we ran, written down, yours to keep and apply to the rest of your estate with or without us.

The guides are the credential. They are not the asset.

We publish the whole method for free because a reader who can tell the difference will find it eventually, and a firm that hides it looks like every other vendor. What is not published is the enforcement: the mapping method, the routing rules, the thresholds and the baselines.

If the two mechanisms above are the part you are missing, that is the conversation. If they are not, the guides are complete on their own and we would rather you started with them.