Decide what a useful result means.

Use the same tasks and review method to separate access, successful responses and answers that help your application.

The included review sheet has no measured results. A successful authenticated WhatsDo search has not been demonstrated with this kit.

Review three things.

Task fit

Does the response identify a suitable place in the requested location and category? Check the actual response against the task.

Field usefulness

Which needed fields are present, missing or uncertain? Do not infer field availability from the questions in a synthetic task.

Evidence quality

Can you trace the information to a source? Confirm permitted use and redaction before sharing a response or derived sample.

Create a review sheet

sh run.sh review --results ./private-results --output review.csv

The runner keeps existing review files. For each HTTP 200 response you inspect, set review_status to reviewed and useful_result to yes or no. Keep unknowns explicit.

Count reviewed outcomes

sh run.sh review --summary review.csv

The summary separates unmeasured tasks, HTTP success, manual review and useful results. None is a claim about general coverage, pricing or production reliability.

Tell us what you are building.

Share the problem your application needs to solve. Access and available capabilities need to be confirmed for your use case.