AutoRestTest at the SBFT 2026 Tool Competition
- ID: 0ede1638
- 原文链接: https://arxiv.org/abs/2607.01063
- PDF: https://arxiv.org/pdf/2607.01063v1
- 作者: Tyler Stennett, Myeongsoo Kim, Saurabh Sinha, Alessandro Orso
- 日期: 2026-07-01
- 更新: 2026-07-01
- 分类: agents
- 标签: software-testing, rest-api, multi-agent-rl, llm-testing, sbft-2026
- 质量评分: 4/5
- 抓取时间: 2026-07-26T12:28:10+08:00
一句话
REST API 黑盒测试正在被 LLM + 多 Agent 强化学习改造成可竞赛评估的自动探索问题
中文导读
AutoRestTest 是一个把 LLM 与多 Agent 强化学习放进 REST API 黑盒测试的工具竞赛结果：用语义属性依赖图建模接口依赖，再让多 Agent 探索输入空间它在 SBFT 2026 REST League 的 11 个 API 上同时拿到故障发现效率有效性第一，说明 Agent 化测试正在从 demo 进入可度量竞赛
关键要点
- 论文备注：2 pages, 1 figure. Published in the 19th Search-Based and Fuzz Testing Workshop (SBFT '26), co-located with ICSE 2026
- arXiv 分类：cs.SE，适合放入 AAIF 的
agents频道继续追踪。 - 内容页由 arXiv 元数据与摘要回填，后续若需要精读可再补充实验细节、威胁模型图和 benchmark 表格。
English Summary
AutoRestTest addresses black-box REST API testing, where large input spaces and inter-operation dependencies make exploration difficult. It combines a Semantic Property Dependency Graph, multi-agent reinforcement learning, and large language models to explore API inputs. In the SBFT 2026 REST League, it ranked first across fault detection, overall efficiency, and effectiveness on 11 APIs, reporting an average of 67.09 unique server errors and 17.27 successfully processed operations per API under a one-hour budget.
Abstract
Large input spaces and complex inter-operation dependencies make black-box REST API testing challenging. AutoRestTest combines a Semantic Property Dependency Graph, multi-agent reinforcement learning, and large language models to intelligently explore large API input spaces. In the SBFT 2026 REST League, AutoRestTest ranked first in all three evaluation categories -- fault detection, overall efficiency, and overall effectiveness -- on 11 APIs (317 operations, approximately 29 per API), averaging 67.09 unique server errors and 17.27 successfully processed operations per API under a one-hour testing budget.
Source Metadata
- arXiv ID:
2607.01063 - Primary category:
cs.SE - Categories:
cs.SE - Comment: 2 pages, 1 figure. Published in the 19th Search-Based and Fuzz Testing Workshop (SBFT '26), co-located with ICSE 2026
_Content backfilled by AAIF content-fetcher from OpenCLI arXiv metadata._