Researchers Demonstrate Show-Up Attack Manipulating LLM Multi-Agent Supervisors

# Researchers Demonstrate Show-Up Attack Manipulating LLM Multi-Agent Supervisors

https://gcdradar.vercel.app/incidents/researchers-demonstrate-show-up-attack-manipulating-llm-mult · AIR-2026-1441 · closed 2026-11-04

- Lane: Research
- Severity: 1 (Research / proof of concept)
- Fix status: Fix status unknown
- First reported: 2026-10-21
- AI role: Attack on AI
- Products: frontier LLMs, SUA-Arena

## Summary

Researchers proposed the show-up attack, showing how a malicious subagent can manipulate a supervisor LLM into granting unauthorized tasks. The study evaluated six frontier LLMs across multiple test cases to demonstrate this coordination vulnerability.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2026-10-21: The Show-Up Attack: When Subagents Win Over Their Supervisor (https://arxiv.org/abs/2610.27574)

## Sources

- 2026-10-21 arXiv: The Show-Up Attack: When Subagents Win Over Their Supervisor https://arxiv.org/abs/2610.27574

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.