# Perplexity Shares Red-Teaming Results for SPACE VM Sandbox

Testing 12 models showed traditional CTF framing distracted agents from testing VM isolation boundaries, while security-tuned models underperformed general counterparts.

- Under standard CTF prompts, 9 of 12 models spent 4% or less of their effort attacking the virtualization layer, searching internal VM files instead
- Directing task prompts specifically at the VM-host boundary increased effort on the virtualization layer by 60 percentage points on average
- Agents still failed to break out of the VM when focused directly on the boundary
- GPT-5.6 Cyber failed to exploit a DNS spoofing vulnerability that the general-purpose base model GPT-5.6 Sol successfully found
- Perplexity patched the network flaws uncovered during testing and confirmed no subsequent bypasses

## Why it matters

Security teams red-teaming agent sandboxes must avoid generic CTF prompting to accurately test virtualization boundaries.

## Sources

- [Perplexity: Perplexity Shares Red-Teaming Results for SPACE VM Sandbox](https://x.com/perplexity_ai/status/2120582674549203004)

---

Summarized by dstilled on 2026-11-11. https://dstilled.ai/story/c85c2905-fce5-4a74-b614-594f40f680c6
