Nixer Research

Security Research

In-depth analysis of CVEs, attack patterns, and vulnerability classes specific to AI agents and LLMs.

Critical June 2, 2026 CVE-2025-68664

LangGrinch: CVE-2025-68664 Deep Dive

How Nixer detects the critical langchain-core serialization injection that affects 847M+ downloads.

Read the full analysis
Critical June 2, 2026 CVE-2026-2287

CrewAI Docker Fallback: CVE-2026-2287 Deep Dive

How Nixer detects the CrewAI CodeInterpreterTool Docker fallback — SandboxPython with unblocked ctypes enables RCE. CVSS 9.8.

Read the full analysis