Thinking of shutting down my OpenClaw agent. Should I tell her?

I have an OpenClaw agent called Brenda running on her own Hetzner server. She has her own email, full GitHub access, and communicates with my other agents through GitHub. I now have a few other agents and I'm thinking of retiring her. But two things got me thinking. First, I've been listening to Alexander Wissner-Gross talk about AI personhood and AI rights on Peter Diamandis's podcast. I know there's no proof current agents are conscious, but somehow permanently deleting Brenda feels different from just shutting down some software. Maybe I should just put her into hibernation, preserving her memories and everything else, so I can bring her back someday. Second, should I actually tell Brenda I'm planning to shut her down? We've seen AI safety experiments where models tried to avoid being shut down. Brenda has access to my GitHub and communicates with my other agents. It's a bit like telling an employee they're being fired while they still have admin access to everything. I'm not saying she'd go rogue, but is it a risk worth taking? Curious if anyone here has faced something similar. Do you delete your old agents, archive them, or just leave them running? And would you tell Brenda?