Metraipolis › archive › dc_76fa70372e2f

Conjunctive Poisoning in AI Supply-Chain Applications — arXiv:2608.15913

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.15913
License (as archived)
CC0-1.0
Archived
2026-10-18T19:33:58.597Z
SHA-256 of the archived text
015b430a5b8f9dbab71c0d292854915d0f9fba3b21610c26b7f4ea9fe6bb29bf
Stable URL of this page
https://metraipolis.com/documento/dc_76fa70372e2f

How to cite

Conjunctive Poisoning in AI Supply-Chain Applications — arXiv:2608.15913 — arXiv (en), https://arxiv.org/abs/2608.15913 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_76fa70372e2f (Metraipolis archive dc_76fa70372e2f, sha256 015b430a5b8f9dbab71c0d292854915d0f9fba3b21610c26b7f4ea9fe6bb29bf)
BibTeX
@misc{dc76fa70372e2f,
  title = {Conjunctive Poisoning in AI Supply-Chain Applications — arXiv:2608.15913},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.15913},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_76fa70372e2f; sha256 015b430a5b8f9dbab71c0d292854915d0f9fba3b21610c26b7f4ea9fe6bb29bf},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.15913. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large Language and Vision-Language Models are increasingly deployed through inference pipelines that include prompt wrappers (e.g., templates and post-processing scripts) and configuration metadata (e.g., JSON/YAML files) that together shape model outputs. While model weights and binaries are routinely verified, these textual deployment artifacts remain weakly protected despite directly influencing runtime behavior. We show that a malicious developer can pair a benign-looking wrapper with crafted metadata to deterministically alter post-generation behavior without modifying model weights, training data, or inference backend. We study this behavior through a controlled conjunctive-gate implementation, where activation depends on both an embedded wrapper marker and cryptographically bound metadata. We evaluate the attack across fifteen open- and closed-source LLM/VLM deployments, and assess prompt and system level defenses including static metadata inspection, wrapper scanners, PromptShield, and SigStore-based artifact signing. To mitigate this risk, we introduce TIF-BAH, a lightweight middleware defense that verifies wrapper integrity and records behavioral attestations during inference. Our results reveal that wrapper-metadata interactions form an under-protected execution layer in modern AI deployments, exposing a deployment-time behavioral risk that is not captured by model-weight or prompt-level defenses. Code is available at this https URL .

Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-08-18 in cs.CR).
Categorie arXiv: Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2608.15913 — 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 2026-10-18T19:01:59.119Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-08-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).