Home / May 12, 2026 / Story
0
#3 The Hacker News general May 10, 2026 at 12:41 UTC

Ollama Out-of-Bounds Read Vulnerability Allows Remote Process Memory Leak

By [email protected] (The Hacker News)

AI Summary

Cyera researchers disclosed CVE-2026-7482 (CVSS 9.1), an out-of-bounds read vulnerability in Ollama dubbed 'Bleeding Llama,' allowing remote unauthenticated attackers to leak the entire process memory of affected servers. The flaw is estimated to impact over 300,000 globally exposed Ollama servers. Given Ollama's widespread use for self-hosted LLM inference, this vulnerability poses significant risk of sensitive data exfiltration from AI workloads.

Relevance score: 88.0/100

# More from May 12