Metraipolis › archive › dc_7317c2cfdb3e

Evidence-Guided Recovery for AI Agents with Ambiguous Tool Outcomes and Partial Execution — arXiv:2610.41453

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.41453
License (as archived)
CC0-1.0
Archived
2026-11-03T07:51:20.921Z
SHA-256 of the archived text
043fca4d2aaf1b04d2a8d621266ddb07b259720fa495f1991253f242c692b6a5
Stable URL of this page
https://metraipolis.com/documento/dc_7317c2cfdb3e

How to cite

Evidence-Guided Recovery for AI Agents with Ambiguous Tool Outcomes and Partial Execution — arXiv:2610.41453 — arXiv (en), https://arxiv.org/abs/2610.41453 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7317c2cfdb3e (Metraipolis archive dc_7317c2cfdb3e, sha256 043fca4d2aaf1b04d2a8d621266ddb07b259720fa495f1991253f242c692b6a5)
BibTeX
@misc{dc7317c2cfdb3e,
  title = {Evidence-Guided Recovery for AI Agents with Ambiguous Tool Outcomes and Partial Execution — arXiv:2610.41453},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41453},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7317c2cfdb3e; sha256 043fca4d2aaf1b04d2a8d621266ddb07b259720fa495f1991253f242c692b6a5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41453. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Tool timeouts can leave AI agents facing ambiguous tool outcomes and partial execution: an action may have failed, committed or remained in flight while other workflow steps have already completed. We formulate evidence-guided task recovery over approved obligations, provider capabilities and states compatible with public evidence. A shared admissibility check handles key expiry, authorization timing, dependencies and residual effects; a terminal certificate separates actual goal satisfaction from publicly supported safe completion. The known-model implementation combines standard observation-contingent planning with a separately implemented effect-history adjudicator. In a finite study of 37 assignments, planning and three active controls each certify 28 completions; four saved synthetic fee units incur additional computation. A preregistered assignment-held-out comparison then evaluates planning advice against evidence prompting with two local model bundles. Across 58 paired units in 19 previously developed task families, advice certifies 30 completions versus 9, a 36.2-percentage-point difference (exact family sign-flip p=0.0078), using 272 proposals versus 399. All 116 trajectories pass independent replay. Both models exhibit a positive completion difference. The comparison measures the evaluated advice interface under a common shield; search-versus-rule advice requires a separate matched comparison. Unknown and handoff outcomes remain explicit, with guarantees tied to the declared provider capabilities rather than arbitrary third-party exactly-once execution.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.41453 — 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 2026-11-03T07:49:31.063Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).