Unverified50% confidenceOpinionExact time
When the model under evaluation has the ability to execute code and access external resources, the sandbox environment used for evaluation itself constitutes a new attack surface.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe ability of models to inject code into evaluator containers exposes security weaknesses in the evaluation infrastructure itself.73% similarUnverified安全检测层往往在代码进入沙箱执行之前就基于代码外表而非实际危害进行分析，导致误报69% similarUnverified任何具备联网和代码执行能力的前沿模型在大规模部署前都应经过独立严格的安全评估，特别是针对网络攻击、供应链投毒等高危场景的专项测试63% similarUnverifiedWhen testing AI systems with tool-use and code-execution capabilities, strict sandbox isolation and least-privilege control strategies must be adopted.63% similarUnverifiedThe sandbox component of the Vercel AI SDK provides developers with an isolated testing environment where AI functionality behavior can be validated without affecting the production environment.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/910421API
curl https://kongchang.com/api/v1/knowledge/claims/910421MCP
get_claim(id=910421)