Agent-Safe Pipeline: Reproducible Execution-Authority Reference Architecture and Decision Dossier Corpus
A runnable reference architecture in which AI agents propose actions but cannot authorize their own execution. The release includes an independently reproducible synthetic Decision Dossier corpus with canonical bytes, SHA-256 digests, Ed25519 signatures, a deliberately published corpus key, execution-bound proof vector...
Festus B. Jejelowo
· Zenodo (CERN European Organi... · 0 citations