Not at all. The sender signs all messages with the same key used to generate their peer ID.
Note: If you’re trying to detect the source IP address for a message, you’ll have to try to find the source peer in the DHT. There’s no guarantee you’ll be able to do that or that the peer will actually exist in the DHT.
What measures can a message sending node take to make sure it will not get into DHT?
In other words, what sender might do to ensure it will not be traced?
So, it’s theoretically possible to use IPFS/pubsub anonymously.
Unfortunately, it’s not just a matter of flipping a switch somewhere.
This is one of those things where you need a complete understanding of
how libp2p/pubsub/ipfs works to implement a correct solution.