Metraipolis › archive › dc_c8aba92a414f
Hearsay: Can an Auditor Trust the Record a Deployed Agent Harness Writes? — arXiv:2609.32495
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.32495
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T21:00:17.586Z
- SHA-256 of the archived text
e2c91bea28966342eb07d28591436a31b9fb117dd9ebe3158dde91262cc9604c- Stable URL of this page
https://metraipolis.com/documento/dc_c8aba92a414f
How to cite
Hearsay: Can an Auditor Trust the Record a Deployed Agent Harness Writes? — arXiv:2609.32495 — arXiv (en), https://arxiv.org/abs/2609.32495 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c8aba92a414f (Metraipolis archive dc_c8aba92a414f, sha256 e2c91bea28966342eb07d28591436a31b9fb117dd9ebe3158dde91262cc9604c)
BibTeX
@misc{dcc8aba92a414f,
title = {Hearsay: Can an Auditor Trust the Record a Deployed Agent Harness Writes? — arXiv:2609.32495},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.32495},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c8aba92a414f; sha256 e2c91bea28966342eb07d28591436a31b9fb117dd9ebe3158dde91262cc9604c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32495. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
An agent harness, the code that turns a model into an agent, writes its own record of each run, and that record is all a later reader gets when a run is disputed, investigated or audited. We call a record evidentiary when a reader who was not there can check it without trusting the writer. Across sixteen deployed frameworks, none writes one in full. Hearsay examines the record, not the task: five harnesses run fourteen tasks, three blinded LLM examiners and a human panel read the records, and every excerpt an examiner quotes is checked mechanically for who wrote it. First, the record lets a reader name the fault but not prove how the run went. Examiners name the right fault in 74 to 91% of 140 runs, but the fault can be proved only from two files the benchmark adds; for what happened in between, fewer than one citation in ten lands on anything the harness did not write, and the examiner with the fewest false alarms catches half of the entries we delete, rewrite or fabricate. Second, the remedy is a second author, not a stronger seal on the first. An append-only log of what passes between harness and model, kept outside the harness and read against the record in both directions, reports all 28 omissions and fabrications we made a harness commit as it ran, where a hash chain over the harness's own record passes all 28. Handed the log, examiners keep their fault verdicts but rest more of their citations on what the harness did not write. What makes a record evidence is who writes it, not what is captured. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI). Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2609.32495 — 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 2026-10-07T20:29:02.353Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).