Netskope Report Warns of AI Data-Access, Agentic Coding and Supply-Chain Risks
Summary
Netskope’s AI report, covering June 2025–July 2026, identifies rising downstream data-access risks from MCP, agentic coding tools as potential execution vectors, and weaponized AI supply-chain threats.
Key points
- MCP traffic increased fourfold in the 10 weeks before the report; the report says connections between internal data stores and external models can expose data users or agents are not authorized to access.
- Claude Code and Codex adoption reached 75% and 58% of organizations, respectively, creating potential execution vectors for malicious code.
- The report describes attackers using trojanized developer tools and fake AI installers, and manipulating AI apps to surface malicious code or links.
- Netskope frames downstream policy violations as the primary architectural risk, calling for granular downstream inspection.
Article Details
- Publisher
- Netskope Research
- Report Period
- June 2025 through July 2026
- Scope
- AI-related security risks, including downstream policy violations, agentic coding apps, and AI supply-chain threats.
- Key Statistics
- MCP traffic increased 4x in the 10 weeks before the report.
- Claude Code adoption was reported in 75% of organizations; the comparison timeframe and measurement details were not specified.
- Codex adoption was reported in 58% of organizations; the comparison timeframe and measurement details were not specified.
- Recommendations
- Use granular downstream inspection to address risks from MCP connections between internal data stores and external models.
Products
Claude CodeThe adoption of agentic coding tools—specifically Claude Code and Codex—has moved from negligible levels a year ago to 75% and 58% of organizations, respectively.CodexThe adoption of agentic coding tools—specifically Claude Code and Codex—has moved from negligible levels a year ago to 75% and 58% of organizations, respectively.