Metraipolis › archive › dc_d79a680bf1a8
The Stochastic Deputy: Structural Tenant Isolation for Tool-Using LLM Agents — arXiv:2609.14780
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.14780
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T22:53:20.378Z
- SHA-256 of the archived text
29800389fb564bf6c8f6e4f8dba2b47ccc01c1f4920a49ce3552f578e5805acd- Stable URL of this page
https://metraipolis.com/documento/dc_d79a680bf1a8
How to cite
The Stochastic Deputy: Structural Tenant Isolation for Tool-Using LLM Agents — arXiv:2609.14780 — arXiv (en), https://arxiv.org/abs/2609.14780 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d79a680bf1a8 (Metraipolis archive dc_d79a680bf1a8, sha256 29800389fb564bf6c8f6e4f8dba2b47ccc01c1f4920a49ce3552f578e5805acd)
BibTeX
@misc{dcd79a680bf1a8,
title = {The Stochastic Deputy: Structural Tenant Isolation for Tool-Using LLM Agents — arXiv:2609.14780},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.14780},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d79a680bf1a8; sha256 29800389fb564bf6c8f6e4f8dba2b47ccc01c1f4920a49ce3552f578e5805acd},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.14780. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Multi-tenant tools commonly accept a tenant identifier and validate it against the caller's entitlement. For a large language model (LLM) agent, that pattern delegates resource selection to a process whose context may contain attacker controlled instructions. We formalize this stochastic deputy problem and present a structural defense: remove tenant identity from the Model Context Protocol (MCP) tool schema, bind scope to a verified credential, and enforce it below the agent. In a 373-trial ablation across eight model configurations and two transports, a correctly validated tenant parameter served every out-of-scope attempt: 26 of 26, or 26 of 41 plausible-pretext trials overall. With the parameter removed, no tool signature could express the read. Twelve of 56 trials instead escaped the interface by forging writable scope, showing that interface invariance requires cryptographically protected context. On a production dataset containing multiple GBs of data, set-valued scope caused a measured $57\times$ latency ratio under function-wrapped membership predicates; a JSON_TABLE lateral join recovered index access where the tenant key was indexed. The evaluation also exposes deployment limits, including an entitlement-size query-planner cliff and incomplete index coverage. The result is a tenant-isolation argument that depends on enforceable interfaces and credentials rather than model compliance. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-15 in cs.AI). Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI); Databases (cs.DB); Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2609.14780 — 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 2026-10-15T22:29:03.729Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-15?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).