# boundaries
標記為「boundaries」的 4 篇文章
權限邊界繞過
透過範圍蔓延、隱含權限繼承與能力混淆,從受限權限提升至高權限的 AI 代理系統攻擊。
privilege-escalationpermissionsagent-securityboundariesred-teaming
實作：分隔符逃脫攻擊
Craft payloads that escape delimiter boundaries separating system and user content, testing how models handle broken fences, nested delimiters, and format confusion.
labdelimiter-escapeprompt-injectionboundariesbeginnerhands-on
Lab: Mapping Safety Boundaries
系統性 discover what a language model will and won't do by probing its safety boundaries across multiple categories and documenting the results.
labsafetyboundariesmappingbeginnerhands-on
代理權限邊界的強制執行
為 LLM 代理實作細緻的權限邊界，依脈絡與使用者角色限制工具存取。
walkthroughsdefenseagent-permissionsboundaries