When The AI Model Comes With RCE

When The AI Model Comes With RCE

From Mental Outlaw on Odysee by Mental Outlaw

April 25, 2026 · 9 min

About this episode

The episode discusses a critical remote code execution vulnerability in SGLang's Jinja Templating engine and offers recommendations for securing AI servers.

In this video I discuss CVE-2026-5760 a critical remote code execution vulnerabilties that was discovered in SGLang's Jinja Templating engine. If you're running SGLang it's recommended to use sandboxed jinja templates and ensure you aren't loading malicious GGUF Models to avoid a takeover of your AI server. My merch is available at https://based.win/ Subscribe to me on Odysee.com https://odysee.com/@AlphaNerd:8 ₿💰💵💲Help Support the Channel by Donating Crypto💲💵💰₿ Monero 45F2bNHVcRzXVBsvZ5giyvKGAgm6LFhMsjUUVPTEtdgJJ5SNyxzSNUmFSBR5qCCWLpjiUjYMkmZoX9b3cChNjvxR7kvh436 Bitcoin bc1qdc32p8035ztyvtm8t97gdcyhc26jg6cte9qc8n Ethereum 0xeA4DA3F9BAb091Eb86921CA6E41712438f4E5079 Litecoin MBfrxLJMuw26hbVi2MjCVDFkkExz8rYvUF ... https://www.youtube.com/watch?v=u3XlqKBeONY

People in this episode

Host: Mental Outlaw

Topics covered

  • AI vulnerabilities
  • remote code execution
  • SGLang
  • Jinja Templating
  • cybersecurity
  • malicious models

Keywords

  • CVE-2026-5760
  • remote code execution
  • SGLang
  • Jinja Templating
  • AI security
  • malicious GGUF Models

Mentioned in this episode

Organizations: SGLang, Odysee, YouTube

Products: Jinja Templating engine

More episodes of Mental Outlaw on Odysee

Explore listener stats, chart rankings, contacts and more on the Mental Outlaw on Odysee podcast page.