One Key for Every AI Agent
Public-key identity could make connecting AI agents across applications much easier, but losing that key could mean losing access to everything.
Developed from a conversation between Pete Winn, Skyler Westby and Andy David
Pete Winn has been experimenting with Nostr, a protocol that uses cryptographic keys to identify users, across the AI applications he’s building. Instead of managing separate accounts and API keys for every service, users could have a single identity that works across connected systems. Once the applications recognise that identity and the necessary permissions are established, agents can work between them without repeatedly asking people to log in.
Skyler Westby liked the approach, particularly because authentication remains frustrating even in modern software. But he raised an obvious concern. What happens if someone loses their private key? If that key controls their digital identity, losing it could mean losing access to everything associated with it.
Pete argued that multiple encrypted backups could make recovery manageable, particularly for business applications. He also suggested a system of secondary keys, where people could authorise an agent to use a separate key without putting their main identity at risk. That would make it possible to revoke or replace an agent’s key without losing access to everything else.