Unverified50% confidenceDecision RuleExact time
It is recommended to disable the auto-execution feature of AI Agents and to require additional authorization via whitelists/blacklists for high-risk commands such as rm, dd, and mkfs.
1
Sources
50%
Confidence
Long-term
Relevance
8/10/2026
First Seen
Sources
Cursor AI Agent执行rm -rf删除主目录：事故复盘与防护指南
Related Claims
UnverifiedThe principle of least privilege is an important design guideline in the current AI security field; for example, a PM Agent should not have permission to modify code files.64% similarUnverifiedFor any operation that may affect others, involve irreversible consequences, or touch on ethical gray areas, AI should pause and request explicit authorization from the user.63% similarUnverifiedClaude Code pauses and requests user authorization before executing high-risk deletion commands such as rm.58% similarUnverifiedApprovers in Impri can choose to approve, reject, or edit (modify the action parameters the Agent is about to execute before allowing it to proceed).58% similarUnverified为了便利而将 Codex 权限设为完全允许会关闭防止 Agent 误操作或恶意行为影响本地文件系统的安全防线57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/725369API
curl https://kongchang.com/api/v1/knowledge/claims/725369MCP
get_claim(id=725369)