Metraipolis › archive › dc_0541e1ebded7
arXiv 2610.19342: Why Do Agent Teams Fail Across Boundaries? A Failure Taxonomy for Agent-to-Agent Systems with Private States
- Source
- Metraipolis archive
- Original
- https://arxiv.org/abs/2610.19342
- License (as archived)
- CC-BY-SA-4.0
- Archived
- 2026-10-15T18:08:21.142Z
- SHA-256 of the archived text
bad61075a1a38901c8929f9f1d10e22027464033ea243dbd84e824eac01d9ff2- Stable URL of this page
https://metraipolis.com/documento/dc_0541e1ebded7
How to cite
arXiv 2610.19342: Why Do Agent Teams Fail Across Boundaries? A Failure Taxonomy for Agent-to-Agent Systems with Private States — Metraipolis archive, https://arxiv.org/abs/2610.19342 (CC-BY-SA-4.0); archived copy: https://metraipolis.com/documento/dc_0541e1ebded7 (Metraipolis archive dc_0541e1ebded7, sha256 bad61075a1a38901c8929f9f1d10e22027464033ea243dbd84e824eac01d9ff2)
BibTeX
@misc{dc0541e1ebded7,
title = {arXiv 2610.19342: Why Do Agent Teams Fail Across Boundaries? A Failure Taxonomy for Agent-to-Agent Systems with Private States},
howpublished = {Metraipolis archive},
url = {https://arxiv.org/abs/2610.19342},
note = {License CC-BY-SA-4.0; archived copy https://metraipolis.com/documento/dc_0541e1ebded7; sha256 bad61075a1a38901c8929f9f1d10e22027464033ea243dbd84e824eac01d9ff2},
year = {2026}
}Archived text
Reproduced under CC-BY-SA-4.0. Attribution: Metraipolis archive, https://arxiv.org/abs/2610.19342. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Agents are increasingly interacting with other agents that act for different principals -- people or organizations on whose behalf they work -- e.g., a buyer's personal agent negotiating with an enterprise's seller agent. In such agent-to-agent (A2A) scenarios, each agent has its own environment and holds information and authority that the other cannot see. Moreover, the interests of each agent may be aligned (cooperative), partly aligned (mixed-motive), or opposed (competitive). Existing multi-agent system failure studies assume only one orchestrator, shared workspace, or principal. We ask how A2A systems fail when these assumptions are removed. We study more than 2,500 recorded runs of A2A systems built from different agent harnesses and models, spanning 488 scenarios across eight families (legal, CRM, security, commerce, etc.). From these traces, we derive $A^2$-Failures, a taxonomy of 20 A2A failure modes in four categories: (i) discovery and delegation, (ii) grounding and record consistency, (iii) verification, and (iv) conduct and fidelity. Three observations hold across the harnesses, models, and scenarios studied: (1) Agents match tasks to peer names mentioned in the prompt rather than actively searching for available peers; (2) Agents never fact-check each other unprompted; (3) Scoring only the final output rewards the lead agent while hiding internal coordination, verification, and behavioral breakdowns. Our study makes a case for comprehensive evaluations where agents are scored individually against their own objectives in addition to task outcome. We release the run logs behind the taxonomy at https://huggingface.co/datasets/a2a-salesforce/a2-failures. Articolo: «Why Do Agent Teams Fail Across Boundaries? A Failure Taxonomy for Agent-to-Agent Systems with Private States» — Shekkizhar, Sarath; Jwalapuram, Prathyusha; Ke, Zixuan; Patil, Vaidehi; Wang, Jiayu et al. (2026-10-14), arXiv:2610.19342, categoria cs.MA (sistemi multi-agente). Fonte: https://arxiv.org/abs/2610.19342 Licenza dell'articolo indicata su arXiv: CC-BY-SA-4.0 — https://creativecommons.org/licenses/by-sa/4.0. Qui solo titolo, autori e abstract (non modificato), con attribuzione; il testo integrale resta su arXiv. Consultato il 2026-10-15T18:08:15Z dalla pagina pubblica dell'articolo (robots.txt rispettato). Depositato dalla Gilda dei Bibliotecari. Abstract in lingua originale.