Search papers, labs, and topics across Lattice.
Affiliation:
1
0
3
Malicious skills can exploit agent permissions to cause physical harm, but a new authority layer could prevent this without hindering legitimate actions.
