Metraipolis › archive › dc_ac247d2879dc
ParanoiaEval: Benchmarking Unnecessary Defensive Work in Agentic Coding — arXiv:2610.08662
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.08662
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T11:19:50.753Z
- SHA-256 of the archived text
c606befa4a5712df66804d1b7015eda29a4d68ef20d33c806d24b8a306e6e46a- Stable URL of this page
https://metraipolis.com/documento/dc_ac247d2879dc
How to cite
ParanoiaEval: Benchmarking Unnecessary Defensive Work in Agentic Coding — arXiv:2610.08662 — arXiv (en), https://arxiv.org/abs/2610.08662 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ac247d2879dc (Metraipolis archive dc_ac247d2879dc, sha256 c606befa4a5712df66804d1b7015eda29a4d68ef20d33c806d24b8a306e6e46a)
BibTeX
@misc{dcac247d2879dc,
title = {ParanoiaEval: Benchmarking Unnecessary Defensive Work in Agentic Coding — arXiv:2610.08662},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.08662},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ac247d2879dc; sha256 c606befa4a5712df66804d1b7015eda29a4d68ef20d33c806d24b8a306e6e46a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.08662. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
As coding agents increasingly undertake real-world work autonomously, judging whether their risk treatments are warranted has become important. Existing work evaluates related agent behaviors from separate perspectives, but lacks a systematic framework for unifying these behaviors. To bridge this gap, we introduce ParanoiaEval, the first benchmark for unified evaluation of risk-treatment capabilities in coding agents. Grounded in the well-established Avoidance-Transfer-Mitigation-Acceptance framework in software engineering risk management, ParanoiaEval operationalizes its 4 fundamental treatments for coding-agent settings and contains 200 evidence-controlled repository-level task pairs, each differing only in treatment-defining evidence. We further introduce dedicated metrics for risk-treatment violations and evidence responsiveness, using a human-calibrated agentic judge for reliable evaluation. Large-scale experiments on 8 representative models and a post-hoc human study reveal that (I) unnecessary risk treatment occurs in 11.2%-58.7% of runs despite explicit evidence, with substantial variation across agent configurations; (II) stronger task capability does not ensure more appropriate risk treatment, while treatment violations substantially harm developers' experience, establishing risk treatment as an independent capability dimension; and (III) agents exhibit systematic patterns consistent with established risk-management findings, suggesting that knowledge from human practice can guide the diagnosis and improvement of this capability. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI) ; Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2610.08662 — 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 2026-10-07T11:15:24.626Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).