Executive Summary
NSFOCUS analyzed prompt-injection data leakage incidents from July to August 2025 involving LLMs connected to Google Drive, SharePoint and GitHub, where chat records, credentials, API keys and business data were exposed.
What Happened
NSFOCUS analyzes multiple real-world prompt injection–driven data leakage incidents involving LLM applications between July and August 2025, including LLMs connected to third-party services like Google Drive, SharePoint, and GitHub.[1] The report describes how attackers used crafted prompts and sharing links to exfiltrate user chat records, credentials, API keys, and confidential business data from integrated SaaS platforms, and it highlights disclosure timelines such as Zenity’s August 6, 2025 vulnerability publication.[1]
Affected Products
- Google Drive
- SharePoint
- GitHub
Vulnerability / Attack Technique
Prompt injection leading to SaaS data leakage
Technical Mechanism
Crafted prompts and sharing links led connected LLMs to expose data from integrated services.
AI Security Category
LLM01:2025 Prompt Injection
Part of the Prompt Injection topic.
Relevant Frameworks
RealGround Analysis
Integrations are where the data is. RealGround analysis: scope each LLM integration to the minimum folders and repos.
Recommended Controls
- Scope LLM integrations to specific folders and repositories
Related RealGround Service
Secure AI Agent Build · or start with the free AI Agent Security Scanner.
Sources
NSFOCUS Security Lab: https://nsfocusglobal.com/prompt-word-injection-an-analysis-of-recent-llm-security-incidents/
Summary and analysis written by RealGround from the source above; last updated 2026-10-18.