AI 挖掘的漏洞报告存在误报 华为终端安全奖励计划要求提交前进行人工分析和验证Bug reports from AI mining have false positives Huawei Device Security Reward Program requires manual analysis and validation before submission


AI free token/AI免费积分
AI 挖掘的漏洞报告存在误报 华为终端安全奖励计划要求提交前进行人工分析和验证
感谢IT之家网友 有鲫雪狐 的线索投递!

IT之家 10 月 14 日消息,华为安全应急响应中心今日发布华为终端安全奖励计划关于后续漏洞报告处置规则的公告。

官方表示,今年收到了大量利用 AI 工具辅助挖掘的漏洞报告,其中大多数报告都给出了 PoC(IT之家注:概念验证,证明某个安全漏洞确实存在且可被触发的代码、脚本、请求数据包或详细复现步骤),但仍不可避免地存在因 AI“幻觉”误导而产生的误报。

为了合理分配有限的审核资源,确保华为终端安全奖励计划审核团队能聚焦于真实有效的漏洞,同时维护健康有序的社区生态,现提出如下要求:

一、漏洞报告要求

对于 AI 辅助或自动化发现的漏洞报告,提交前务必开展人工分析和验证。笔者鼓励白帽子给出完整的 EXP 攻击链验证,请确保报告中给出漏洞实际安全影响的证明,涵盖但不限于:

二、不规范行为

白帽子在参与奖励计划的全过程中,应始终保持相互尊重,以专业、友好的方式和语气交流。禁止以下行为:

三、不规范行为处置

另外,官方还调整了工具类漏洞受理范围:

自 2026 年 9 月 30 日起,暂不受理华为终端工具(如 DevEco 等)的漏洞。但若您发现了可导致自研工具公有仓供应链投毒类漏洞或利用该工具可影响华为终端设备安全的漏洞,仍可提交,笔者将依据漏洞影响等因素综合评估后予以答复。

本文来源:https://www.ithome.com/1/013/049.htm
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。

IT Home October 14th news, Huawei Security Emergency Response Center today released the Huawei terminal security reward program on the subsequent vulnerability report disposal rules announcement.

Officials said they received a large number of bug reports this year using AI tools to assist in mining, most of which gave PoCs (IT Home Note: Proof of Concept, code that proves that a security vulnerability does exist and can be triggered, scripts, request packets, or detailed reproduction steps), but there are still inevitably false positives due to AI "hallucinations" misleading.

In order to rationally allocate limited audit resources and ensure that the Huawei Endpoint Security Reward Program audit team can focus on real and effective vulnerabilities, while maintaining a healthy and orderly community ecology, the following requirements are put forward:

AI-assisted or automated bug reports must be manually analyzed and validated prior to submission. The author encourages the white hat to give a complete verification of the exp attack chain. Please ensure that the report provides proof of the actual security impact of the vulnerability, including but not limited to:

The white hat should always respect each other and communicate in a professional, friendly manner and tone throughout your participation in the rewards program. The following are prohibited:

Bug reports from AI mining have false positives Huawei Device Security Reward Program requires manual analysis and validation before submission

In addition, the official also adjusted the acceptance scope of tool bugs:

From September 30, 2026, vulnerabilities in Huawei terminal tools (such as DevEco, etc.) will not be accepted for the time being. However, if you find a vulnerability that may lead to the poisoning of the public warehouse supply chain of the self-developed tool or the use of the tool may affect the security of Huawei's terminal equipment, you can still submit it. The author will respond after comprehensive evaluation based on factors such as the impact of the vulnerability.

This article is from: https://www.ithome.com/1/013/049.htm For copyright issues, please write to: 13951431913#139.com or submit a review on the right.

暂无评论,快来发表第一条评论吧!No comments yet. Be the first!

📮 需求咨询📮 Inquiry