Hey, I looked through the open feature requests on https://jira.atlassian.com/secure/Dashboard.jspa and I couldn't find anything like this - but maybe it's there.
As an admin, I can currently allow or disallow users to generate personal API tokens via authentication policies. That is, I can turn it on or off, and I can limit how many days before the token expires.
Additionally, when users create these tokens, they can choose the scopes.
What I would love to see is the ability for me as admin to limit the scopes my users have available to them.
With the advent of generative AI tools, it's much easier to use the API, and I'm worried about well-meaning users accidentally deleting or modifying data. I would prefer to limit the majority of people to read-only scopes.
Has anyone else found a workaround, or do you see this as a feature request we can upvote?
We recently split our authentication policies so most users are blocked from creating API tokens, so that they have to request being moved to a new policy that allows them to use the API. In our efforts to have better knowledge and awareness of what they are doing, we've found some pages which we were previously unaware of (they might be recent additions...)
https://admin.atlassian.com/o/ORG_KEY/api-tokens
They have a few articles they link to there: https://support.atlassian.com/organization-administration/docs/track-user-api-token-usage-in-your-organization/
Also linking to Automation which allows for the Schedule revocation of stale API tokens of configured inactivity period (there's one for keys too). So it's a step in the right direction, but as far as I'm aware it's only inactivity that you can action easily now.
I would definitely support such a feature request.
It feels like a persistent gap in the Atlassian permissions model, where there are global level permissions that you cannot maintain granular control of, e.g. bulk permissions. Just these blanket all or nothing choices. It's unfortunate.
I'm with @Shanelle Boluyt on this one. The lines between UI and AI usage are being blurred anyway. So even if you could restrict API scopes, the person's account still has the means to perform the actions. Your energy would be better spent focusing on the operations and governance around deletion. Eg:
I am in favour of having more options as an Org admin though - so I still think you should submit a ticket. At least gives Atlassian a data point and they may end up approaching this risk from a different angle with AI tools.