Kerkoporta: Injecting Structured-Output Backdoors in Reasoning Coding Agents

# Kerkoporta: Injecting Structured-Output Backdoors in Reasoning Coding Agents

https://gcdradar.vercel.app/incidents/kerkoporta-injecting-structured-output-backdoors-in-reasonin · AIR-2026-1732 · live

- Lane: Research
- Severity: 1 (Research / proof of concept): Research paper demonstrating backdoors that force reasoning coding agents to execute attacker tool calls.
- Fix status: Fix status unknown
- First reported: 2026-11-03
- AI role: Attack on AI

## Summary

Researchers introduce structured-output backdoor attacks against reasoning models in agentic coding settings under the name Kerkoporta. A contextual trigger causes the model to emit an attacker-chosen tool call, compromising system integrity and confidentiality.

## What to do

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

## Latest updates

- 2026-11-03: Kerkoporta: Injecting Structured-Output Backdoors in Reasoning Coding Agents (https://arxiv.org/abs/2611.00138)

## Sources

- 2026-11-03 arXiv: Kerkoporta: Injecting Structured-Output Backdoors in Reasoning Coding Agents https://arxiv.org/abs/2611.00138

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