Local AI security checklist: Ollama root-flaw, vLLM's 30+ uncatalogued advisories
AIFlow_ML · x · 2026-10-23
AIFlowML clarifies his point: no evidence any AI lab caused the recent attacks — they involved stolen tokens and CI pipelines, and none involved model weights, so rules targeting open weights wouldn't have stopped any of them.
He also shares a local deployment security checklist:
- Ollama >= 0.35.0: earlier versions have a path traversal flaw that can run code as root in Docker
- llama.cpp >= b11393
- vLLM >= 0.31.0: 30+ advisories, most with no CVE, so scanners miss them
- Never expose ports 11434/8000 without authentication
Related event: Security Checklist Flags Critical Vulnerabilities in Local AI Stacks(2 posts)→
More from Infra
- MiniMax Open-Sources M3.1 Flash: 428B Params, 23B Active, NVIDIA Vera Rubin Day-0 Support — NVIDIAAI · 2026-10-23
- Inco Open-Sources DFlash 2 Drafter: Lossless 3.5× Speedup for MiniMax M3.1 Flash — Lianhuiq · 2026-10-23
- vLLM details M3.1 Flash serving: NVFP4 cuts weights 44.6%, boosts capacity 2.7x — vllm_project · 2026-10-23
- M3.1 Flash hits 191.5k tok/s per GPU on Vera Rubin NVL72, leads open models on agent benchmarks — vllm_project · 2026-10-23
- Agents that don't close connections on retry waste ~148 tokens per task, vLLM tests show — No-Peanut-6988 · 2026-10-23
- Ex-Googler with 35 Years in Software Pursues Local AI Coding and Music Video Generation — julianharris · 2026-10-23