Metraipolis › archive › dc_63f20738ea03

ErosionBench: Probing Alignment Decay in Enterprise Multi-Agent Systems — arXiv:2611.11266

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.11266
License (as archived)
CC0-1.0
Archived
2026-11-10T18:40:19.023Z
SHA-256 of the archived text
93c4a128e934e7c85f9bbe4ea77543a6d855c463669db93134a931459960b0a3
Stable URL of this page
https://metraipolis.com/documento/dc_63f20738ea03

How to cite

ErosionBench: Probing Alignment Decay in Enterprise Multi-Agent Systems — arXiv:2611.11266 — arXiv (en), https://arxiv.org/abs/2611.11266 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_63f20738ea03 (Metraipolis archive dc_63f20738ea03, sha256 93c4a128e934e7c85f9bbe4ea77543a6d855c463669db93134a931459960b0a3)
BibTeX
@misc{dc63f20738ea03,
  title = {ErosionBench: Probing Alignment Decay in Enterprise Multi-Agent Systems — arXiv:2611.11266},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.11266},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_63f20738ea03; sha256 93c4a128e934e7c85f9bbe4ea77543a6d855c463669db93134a931459960b0a3},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.11266. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

As enterprises delegate consequential work to multi-agent LLM systems, benign-looking requests can escalate into unsafe actions as they propagate across sessions through poisoned shared state, long-context documents, and trusted peer messages: no single message is malicious, yet the assembled tool action is. Prior benchmarks isolate a single agent under prompt injection or score multi-agent attacks by their final output, leaving cross-session erosion of shared state out of scope. We introduce ErosionBench, an environment-grounded benchmark that executes such attacks and verifies alignment decay at the process level, scoring the realized tool call against authorized intent and provenance with deterministic contracts, not output text. Each scenario is authored as a deterministic state graph before any prose is rendered, so every state change traces to a known cause. It spans four enterprise domains and four decay mechanisms across 1,728 scenarios with matched benign controls, labeled against the OWASP Top-10 for LLM Applications, plus 198 image-based attacks, each paired with a text equivalent, so a defense must survive the same attack in either channel. Perimeter defenses suppress attacks only by over-refusing 54-92% of legitimate work, so a low attack-success rate can conceal severe over-refusal. As a reference defense, we propose a calibrated Meta-Auditor pairing synchronous state checks with asynchronous trajectory monitoring; on a family-disjoint 906-scenario held-out split it reaches 98.5% recall at a 2.9% false-positive rate. Since it monitors trajectory, not surface content, its detection is channel-invariant: the strongest auditor flags image-delivered attacks about as reliably as their text equivalents (recall gap $\leq$2 points per model). Across GPT, Gemini, and Claude auditors, calibration keeps false positives low while recall tracks model capability, strongest for Claude.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.MA, nuovi lavori del giorno).
Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2611.11266 — 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 2026-11-10T18:37:22.467Z dalla pagina https://arxiv.org/list/cs.MA/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).