A worked example of what the site audit produces, so you can see the shape before you install anything.
- The site is made up. It is the demo site in this repository, which has faults planted on purpose.
- The evidence is real script output. The files in
evidence/were written by the five check scripts run against the demo site on a local server.findings.jsonpassesfindings.py check, and the scores in the summary are whatfindings.py scoreprinted. - It shows the headless half of an audit only. No browser walk, accessibility scan or performance run was done, and the report says so in its coverage table.
- Screenshots sit under the findings they support, which is the rule for every report. Here they are the captures the viewport script saves; a full audit adds the walk's own.
| File | What it is |
|---|---|
| summary.md | The one-page, plain-language version |
| report.md | The full report |
findings.json |
The same findings in machine-readable form, used to compare runs |
evidence/ |
Raw output from each script |
screenshots/ |
The images embedded in the report |
To reproduce the evidence, follow "Try them on the demo site" in
docs/scripts.md and pass --today 2026-07-15 --timezone Europe/London to
copy_freshness.py and --mask "Brindlemere,Oakhaven" to similarity.py.