Help
Reading your report
A report reads as the story of one purchase attempt: four stages in the order the agent met them, each either passed or stalled, with the evidence for the verdict attached. Findings are ordered by where they occurred rather than by severity, because a failure early in the journey makes everything after it untested rather than fine.
A report is the narrative of one attempt, annotated with the evidence for each judgement. It is deliberately not a score with a breakdown underneath: the order things happened in is most of the meaning.
Stages come in the order the agent met them
Discover, evaluate, compare, buy. Each is passed or stalled. A stage after a stall is not a failure and not a pass, it is untested, and it is shown as untested rather than filled in.
This is why the first stall is usually the only one worth acting on immediately. Everything below it is a guess until it is cleared.
Findings name a mechanism
A finding is not "your product data is weak". It is the specific thing that stopped the agent, linked to a page explaining the mechanism and the change that removes it.
That is the "forward this to your developer" test: a finding that cannot be handed to somebody who can fix it has not done its job.
Verdicts are stamped, not softened
A stage passed or it stalled. There is no partial credit and no encouraging language over a failure, because the value of the report is that it is checkable rather than reassuring.
What to do first
Fix in the order the agent met them, not in the order they look easy. See the agentic readiness checklist for the sequence.
Questions
Why is a later stage blank?
Because the attempt ended before it. A store that could not be found is not then assessed on its checkout, and reporting a pass there would be inventing a result.
What does a stall mean exactly?
That the agent could not continue, and why. Each stall names a mechanism rather than a symptom, so it points at a change you can make rather than at a score you cannot act on.