Metraipolis › archive › dc_85a28bc77a72
Outcome Monitors: Recovery Affordances for Silent Tool Failures — arXiv:2608.19303
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.19303
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T11:39:24.880Z
- SHA-256 of the archived text
a392117b9245c33d0cc7deb7039b469e6380fca5fa4bbf5f89f9705b98cbdbb1- Stable URL of this page
https://metraipolis.com/documento/dc_85a28bc77a72
How to cite
Outcome Monitors: Recovery Affordances for Silent Tool Failures — arXiv:2608.19303 — arXiv (en), https://arxiv.org/abs/2608.19303 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_85a28bc77a72 (Metraipolis archive dc_85a28bc77a72, sha256 a392117b9245c33d0cc7deb7039b469e6380fca5fa4bbf5f89f9705b98cbdbb1)
BibTeX
@misc{dc85a28bc77a72,
title = {Outcome Monitors: Recovery Affordances for Silent Tool Failures — arXiv:2608.19303},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.19303},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_85a28bc77a72; sha256 a392117b9245c33d0cc7deb7039b469e6380fca5fa4bbf5f89f9705b98cbdbb1},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.19303. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
When a tool call times out, the agent sees the failure and can route around it. A cached error page or negative price can instead arrive in the expected format and be consumed as fact. We introduce Outcome Monitors, which detect violations of outcome contracts mined from task-disjoint traces or derived from public schemas. On a violation, the monitor preserves the result and issues a nonbinding receipt naming the violated property and public recovery tools. In frozen, prespecified evaluations with injected failures, Outcome Monitors raise ToolMaze completion from 10.9% to 28.1% across four models in two provider families and replicate in a third. In tau-bench retail, completion improves by 14.0 and 12.0 points on two tiers. In separate ToolMaze controls, removing the recovery-tool list eliminates the measured gain and restoring it recovers the effect; diagnostic detail and timing produce no detectable differences. Gains concentrate where the fault blocks completion. On a suite transcribed from a published incident taxonomy, detection outside the mined vocabulary falls to 46%, though delivery continues and completion is unchanged. Recovery tools are the active receipt content in these controls; extending detection beyond the contract vocabulary remains open. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-21 in cs.CL). Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2608.19303 — 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 2026-10-18T11:38:27.938Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-21?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).