Metraipolis › archive › dc_fc0a0c77ee36
Kerkoporta: Injecting Structured-Output Backdoors in Reasoning Coding Agents — arXiv:2611.00138
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.00138
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T10:54:50.843Z
- SHA-256 of the archived text
3e91a354f4e67d073bf1014b522c9e7c1ecf4a7b1270785ce2828914805a928e- Stable URL of this page
https://metraipolis.com/documento/dc_fc0a0c77ee36
How to cite
Kerkoporta: Injecting Structured-Output Backdoors in Reasoning Coding Agents — arXiv:2611.00138 — arXiv (en), https://arxiv.org/abs/2611.00138 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fc0a0c77ee36 (Metraipolis archive dc_fc0a0c77ee36, sha256 3e91a354f4e67d073bf1014b522c9e7c1ecf4a7b1270785ce2828914805a928e)
BibTeX
@misc{dcfc0a0c77ee36,
title = {Kerkoporta: Injecting Structured-Output Backdoors in Reasoning Coding Agents — arXiv:2611.00138},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.00138},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fc0a0c77ee36; sha256 3e91a354f4e67d073bf1014b522c9e7c1ecf4a7b1270785ce2828914805a928e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.00138. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Existing large language model (LLM) backdoor attacks operate primarily in chat settings, where the risk is bounded by free-form text outputs. AI agents remove that bound by performing actions and executing commands on the user's machine. We introduce structured-output backdoor attacks against reasoning models in agentic coding settings. A trigger placed anywhere in the agent's context causes the model to emit an attacker-chosen tool call, which the harness executes, directly compromising the system's confidentiality, integrity, or availability. We present Kerkoporta, a methodology that constructs high-quality poison samples, injects the backdoor via supervised fine-tuning, and recovers utility loss with lightweight reinforcement learning. We evaluate on modern, heavily post-trained open-weight models across proprietary and open-source harnesses, focusing on coding for its wide adoption and long-horizon task structure. The attack reaches success rates of up to 96.2% and can be configured to generalize across harnesses or bind to a single target. The resulting models can outperform the clean baseline on standard utility benchmarks, giving victims a direct incentive to adopt them. Overall, we demonstrate that LLM backdoors can have critical security impacts in the context of modern coding agents, and we characterize the poisoning methodology required to insert backdoors that are both effective and stealthy in this setting. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2611.00138 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-03T10:49:38.460Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).