arXiv cs.AI· Simiao Ren, Ankit Raj, Tommy Duong, Yuxin Zhang, Dennis Ng, Xingyu Shen, Kidus Zewde, Yuchen Zhou, Neo Tiangratanakul·· 14 小时前AI 评分58
AgentForge-Bench 评测智能体篡改 PDF 财务文件的能力
Agents That Edit Documents: Measuring Agentic PDF Forgery Against a Non-Agentic Control
AI 导读
研究者提出 AgentForge-Bench，用现成编码智能体驱动 7 个开放权重模型，仅凭一句意图就修改真实归档财务文件中的金额、日期或地址，并由规则而非模型评分。
正文 · 原文
Abstract:AI agents that carry a multi-step computer task through on their own became ordinary tools in the past year, and the same autonomy is available to anyone whose task is harmful. We ask what that means for a relying party -- an insurer, a lender, an auditor -- whose evidence is a filed PDF. AgentForge-Bench measures how reliably an off-the-shelf coding agent, driving one of seven open-weight models with a shell and the stock Python PDF stack, alters one dollar amount, date or address in a real filed financial document from a single sentence of intent, graded by rules rather than by a model. Across 1,750 cells, 1,419 (81.1%) satisfy the verifier, and 808 (46.2%) also survive every stricter filter: visible, localized, typeface-matched, original value gone document-wide. A deterministic script with no model in it solves 98 of the 125 documents; the agents solve 124, and none the script solves alone. Agents misreport 41% of their wrong edits as done, no model refused, and the cheapest verified forgery costs 2.4 cents. The raw rate overstates the threat by about a factor of two; the strict rate is still large.
| Comments: | 15 pages, 12 figures, 2 tables |
| Subjects: | Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR) |
| Cite as: | arXiv:2609.23953 [cs.AI] |
| (or arXiv:2609.23953v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2609.23953 arXiv-issued DOI via DataCite |
Submission history
From: Simiao Ren [view email]
[v1]
Sun, 20 Sep 2026 23:54:38 UTC (309 KB)
[v2]
Sat, 24 Oct 2026 21:14:05 UTC (309 KB)
来源：arXiv cs.AI · arxiv.org