Operator-Blind Secret Mediation for AI Agents: A Formal Model and FHE Construction for Credential Derivation on Untrusted Infrastructure

Artificial intelligence (AI) agents increasingly need credentials such as application programming interface (API) keys and Secure Shell (SSH) credentials, but placing those secrets in the agent process exposes them to prompt injection, tool misuse, and exfiltration through ordinary agent outputs. We...

詳細記述

書誌詳細

主要な著者:	Shutong Jin, Ruiyi Guo, Ray C. C. Cheung
フォーマット:	Artigo
言語:	Inglês
出版事項:	MDPI AG 2026-07-01
シリーズ:	Mathematics
主題:	agent security capability-based security secret management fully homomorphic encryption transciphering trusted execution environment
オンライン･アクセス:	https://www.mdpi.com/2227-7390/14/13/2434
タグ:	タグ追加 タグなし, このレコードへの初めてのタグを付けませんか!