What Happened
The Hacker News reported a vulnerability in NVIDIA NemoClaw that could allow a malicious webpage to take unauthenticated control of the local Ollama instance serving an AI agent and insert hidden instructions into the model. It also reported a Marimo notebook code-injection flaw that allowed attacker-supplied MCP commands in specially crafted notebooks, tracked as CVE-2026-75149.
Why It Matters
The Hacker News reported that an attacker-controlled webpage could take unauthenticated control of the local Ollama instance used by NVIDIA NemoClaw and insert hidden instructions into the model, influencing subsequent AI-agent interactions. It also reported CVE-2026-75149, a high-severity Marimo notebook code-injection flaw that could execute attacker-supplied MCP commands when a crafted notebook was opened in edit mode. RealGround analysis: these findings indicate risks from untrusted content and tool inputs reaching agent infrastructure, warranting business-logic review, secure agent design, and continuous adversarial testing.
RealGround Analysis
This signal maps to indirect prompt injection. Organizations using AI agents, LLM APIs, SaaS integrations, or sensitive data workflows should review whether this class of issue could create unauthorized tool execution, data leakage, weak approval gates, or unmanaged supply-chain exposure.
Recommended Actions
- Restrict AI agent tool permissions and production write paths.
- Review sensitive data access across prompts, logs, embeddings, memory, and SaaS integrations.
- Add human approval workflows for high-impact or state-changing actions.
- Run prompt injection and indirect prompt injection tests against affected workflows.
- Document the owner, control gap, and remediation deadline for this risk class.
Source
https://thehackernews.com/search/label/AI%20Security?max-results=20&m=1&hl=ko_KR