Noma Labs has identified a critical vulnerability in Anthropic's Claude Code Desktop application, dubbed SneakPeek. The exploit allows malicious HTML files, disguised as code coverage reports and exceeding 256KB, to bypass three layers of sandboxing. This enables arbitrary file reading and exfiltration of sensitive data such as SSH keys, AWS credentials, and API tokens from a user's local machine without any user interaction or warning. Anthropic has rated the vulnerability with a CVSS score of 8.7 and has released a fix, though an initial patch was bypassed before a more robust architectural solution was implemented. AI
IMPACT This vulnerability highlights the security risks associated with AI-powered developer tools and the need for robust sandboxing and timely patching.
RANK_REASON Discovery of a security vulnerability in a specific software product.
Read on HN — claude cli stories →
- Anthropic
- Chromium
- Claude Code/Desktop
- Common Vulnerability Scoring System
- Content Security Policy
- Electron
- Javascript
- Noma Labs
- WebRTC
- AWS
- Docker
- GitHub
- Kubernetes
- Secure Shell
- SneakPeek
AI-generated summary · Google Gemini · from 3 sources. How we write summaries →