I have implemented a function that replicates the behavior of the ssh-add -D command.

A button has been added directly in the entry and a second one is placed in the Tools menu.

I see at least three use cases:

• Clearing the agent before adding a new key, this prevents ssh from testing unused keys.
• Ensuring that the agent no longer contains any keys before using agent forwarding.
• Clearing the agent before leaving the workstation without having to open a terminal.

I chose to indiscriminately delete all keys, @darrellenns suggested deleting only the keys that Keepassxc manages.

Switching between these two approaches is rather simple, just by changing the slot in a single connect function.
It is possible to let the user choose the approach. I was thinking of eventually adding a checkBox in the agent configuration page to toggle between the two modes.

Screenshot:

Neat! Flush is kind of a weird word, though. Maybe stick with "Clear Agent".

My initial intention with flush was to emphasize that key deletion was done indiscriminately in the agent.
Similar to flushing a DNS cache, starting fresh.

But indeed, using the term clear is more elegant.
I can modify all instances of flush to clear.

What else needs to be done before proposing a PR?

Looks good otherwise, just make sure the Tools menu item is hidden if ssh agent is disabled.

I made two commits on my repo:

• The first one fixes the button visibility issue.
• The second one replaces occurrences of the word flush with clear.