StealthyPoison: Manipulating LLM Agent Behavior through Procedural Poisoning of Third-Party Skills

# StealthyPoison: Manipulating LLM Agent Behavior through Procedural Poisoning of Third-Party Skills

https://gcdradar.vercel.app/incidents/stealthypoison-manipulating-llm-agent-behavior-through-proce · AIR-2026-1888 · live

- Lane: Research
- Severity: 1 (Research / proof of concept): Demonstrated research attack with no real-world exploitation reported.
- Fix status: Fix status unknown
- First reported: 2026-11-10
- AI role: Attack on AI
- Products: LLM Agent

## Summary

Researchers presented StealthyPoison, a procedural poisoning technique that corrupts decision-critical rules in third-party skills. The attack induces incorrect task outcomes while preserving advertised functionality and normal execution behavior.

## What to do

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

## Latest updates

- 2026-11-10: StealthyPoison: Manipulating LLM Agent Behavior through Procedural Poisoning of Third-Party Skills (https://arxiv.org/abs/2611.12155)

## Sources

- 2026-11-10 arXiv: StealthyPoison: Manipulating LLM Agent Behavior through Procedural Poisoning of Third-Party Skills https://arxiv.org/abs/2611.12155

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