Executive Summary
ESET reports that as SMBs adopt generative AI and agents, misconfiguration, prompt injection and malicious skills increase the risk of data leaking through AI; many lack AI usage policies.
What Happened
ESETは、中小企業で生成AIやAIエージェントの導入が進む一方、設定不備やプロンプト注入攻撃、不正スキルにより機密情報がAI経由で流出する危険が増していると報告した。[4] 調査では、AI利用規定が未整備の企業が多く、顧客情報や財務データを安易に入力することによるデータ漏洩や権限管理の不備が問題として挙げられている。[4]
Vulnerability / Attack Technique
Misconfiguration, prompt injection and malicious skills
Technical Mechanism
The report describes risk sources rather than one exploit.
AI Security Category
LLM02:2025 Sensitive Information Disclosure
Relevant Frameworks
RealGround Analysis
Malicious agent skills are a supply-chain risk for small teams. RealGround analysis: vet agent plugins and skills before installing.
Recommended Controls
- Adopt an AI usage policy
- Vet agent skills and plugins before installation
Related RealGround Service
AI Security Readiness Assessment · or start with the free AI Agent Security Scanner.
Sources
ITmedia エンタープライズ: https://www.itmedia.co.jp/enterprise/articles/2606/12/news038.html
Summary and analysis written by RealGround from the source above; last updated 2026-10-18.