Methodology
Observe the public path. Preserve the evidence boundary.
The method is intentionally narrow. It checks what a candidate can reach through public pages and avoids claims the available evidence cannot support.
Public evidence only
Audits are limited to public HTTP and HTTPS pages. Credentialed URLs, local addresses, and unsafe private or reserved network destinations are blocked before requests.
Four-stage review
1. Validate the input
Confirm that each target is a permitted public address.
2. Follow relevant public paths
Review the job page, public Apply action, redirects, and reachable ATS destination.
3. Classify the observable state
Separate healthy paths, closed roles, missing actions, broken destinations, login dead ends, unreachable pages, and human-review cases.
4. Produce reviewable evidence
Return an exception-first summary, result table, redirect evidence, and CSV export.
Result states are operational signals
A result helps a team decide what to inspect next. It is not a certification, a complete accessibility review, or proof of why an individual candidate continued or abandoned an application.
Known limits
Dynamic page behavior, bot controls, geography, authentication, vendor changes, and transient network conditions can affect what is publicly observable. Human review remains important for ambiguous results.