Unverified50% confidenceSolutionExact time
AT&T团队对Agent可用的MCP工具进行过滤，仅允许调用经过安全审核的子集以避免数据过度暴露
1
Sources
50%
Confidence
Long-term
Relevance
8/23/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedThe agent using only MCP tools did not set security flags and created a view with a data leakage risk, while the MCP+Skill agent correctly identified and implemented the secure configuration.72% similarUnverified在MCP层可以注入恶意工具返回值、拦截审计Agent的工具调用、模拟工具失败等边界情况，使MCP层成为对抗测试的注入点与观测窗66% similarUnverifiedMCP provides a unified security foundation for Agents within enterprises, defining the scope of permission grants and data access rules.66% similarUnverifiedMCP的设计中Server是被动的，不思考、不决策，只暴露工具等着被调用66% similarVerifiedMCP tools are essentially arbitrary code execution, indirect prompt injection is a real-world threat, and attack vectors can be hidden in resource content rather than in direct user input.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/791255API
curl https://kongchang.com/api/v1/knowledge/claims/791255MCP
get_claim(id=791255)