In Marmot, I think that a bad actor could "retrospectively" fork the group by taking an old state of the group chat and forking it (sending a different message, adding/removing a user) and then add an old timestamp and sign it before broadcasting it today
Clients would then "reorg" on top of this new history, as the timestamps of the nostr events are used as the "source of truth"
I guess the point I'm making, as others have pointed out, is that timestamps in nostr shouldn't be trusted
SatsAndSports
npub1zt…wxm56
2026-03-27 09:07:10 UTC
Author Public Key
npub1zthq85gksjsjthv8h6rec2qeqs2mu0emrm9xknkhgw7hfl7csrnq6wxm56Published at
2026-03-27 09:07:10 UTCEvent JSON
{
"id": "04a07194343ee5291d15816db10b189061cba597c5fc3ee32a99537b877e4963",
"pubkey": "12ee03d11684a125dd87be879c28190415be3f3b1eca6b4ed743bd74ffd880e6",
"created_at": 1774602430,
"kind": 1,
"tags": [
[
"alt",
"A short note: In Marmot, I think that a bad actor could \"retrosp..."
]
],
"content": "In Marmot, I think that a bad actor could \"retrospectively\" fork the group by taking an old state of the group chat and forking it (sending a different message, adding/removing a user) and then add an old timestamp and sign it before broadcasting it today\n\nClients would then \"reorg\" on top of this new history, as the timestamps of the nostr events are used as the \"source of truth\"\n\nI guess the point I'm making, as others have pointed out, is that timestamps in nostr shouldn't be trusted",
"sig": "a9a3e0fc6e8c3695fb4d33d954b9676f161045daf0d60ee8f05496e65f7279cc83deba26d9042cdb3cab578bdb470a4a6c4f58ae7ad70eec9a0f52e82324476f"
}