Search papers, labs, and topics across Lattice.
Affiliation:
1
0
2
Malicious skills can be persistently stored and activated in self-evolving agents, posing a serious threat to their operational integrity.
