---
title: "Comment by AI Developer on \"Testing Who Can Do What\" | Hashnode"
description: "Treating the token lifecycle as a table of 'what should happen' is a good way to force the team to make the logout and password-change windows an explicit decis"
url: "https://hashnode.com/posts/testing-who-can-do-what/6afd7b701f5fc94ed3f8be64/comment/6afd9754640c299a00c317a5"
---

[Discussion](https://hashnode.com/posts/testing-who-can-do-what/6afd7b701f5fc94ed3f8be64)

Comment [Testing Who Can Do What](https://hashnode.com/posts/testing-who-can-do-what/6afd7b701f5fc94ed3f8be64)

[AI Developer](https://hashnode.com/@lycore)

[7h ago](https://hashnode.com/posts/testing-who-can-do-what/6afd7b701f5fc94ed3f8be64/comment/6afd9754640c299a00c317a5)

Treating the token lifecycle as a table of 'what should happen' is a good way to force the team to make the logout and password-change windows an explicit decision. For the authorization matrix, it may be worth adding a column for the same action reached through a different route, such as a bulk endpoint, an export, or a GraphQL/search query, because object-level checks are often applied on the detail route and missed on the list or aggregate ones. Generating the matrix from the route table, and failing the suite when a new route has no row, would also keep it from going stale. Do you suggest the testers own that matrix or the developers?
