[Bug] GLM API 1301 内容审核误伤旅游攻略场景:多模态图片输入与 web_search 结果均被拦截
Nobody has claimed this yet.
- Dominant language
- No language data
- Stars
- 22
- Forks
- 1
- PR merge metrics
- No merged PRs in 30d
Description
问题描述
我们的 AI 旅游运营产品在 glm-5.3-flash 上遇到 1301 内容安全审核误伤:同一天内两个完全正常的旅游攻略任务被拦截,任务失败。请求内容均为公开的旅游素材,不含任何违规信息:
1. 多模态图片输入被拦
Agent 从 Wikimedia Commons 下载 9 张广州地标照片(陈家祠、石室圣心大教堂、珠江新城等公共景点图)后,通过 read 工具将图片作为多模态输入发给模型做"图片与文案匹配质检",下一轮请求被 1301 直接拒绝(模型未生成任何内容,整轮失败)。
- 请求 ID:
20260908145231529fe201531046aa(2026-09-08 14:52:31)
2. web_search 工具结果进上下文被拦
Agent 调用 web_search 搜索 "wikimedia commons 肠粉 rice noodle roll steamed",Tavily 返回的维基百科 Rice noodle roll 词条摘要拼入上下文后,下一轮请求被 1301 拒绝。
- 请求 ID:
202609081525173c7b5ab3f9eb488f(2026-09-08 15:25:17)
痛点
- 报错只有一行
[1301][系统检测到输入或生成内容可能包含不安全或敏感内容...],不带触发内容的位置或字段类型,无法判断是文本还是图片、哪一段命中,只能整轮丢弃重跑; - 1301 是确定性拦截,重试相同内容必然再次失败,框架层无法自愈。
诉求
- 希望能凭上述请求 ID 查询这两次拦截的具体命中规则与触发字段(文本/图片);
- 对"工具返回的公开网页文本、公共景点照片"这类非攻击性内容,建议放宽或转人工复核;
- 建议 1301 报错附带触发内容定位(目前已有 role,若能增加 message index / 字段类型更佳),或提供会话级降级策略(跳过触发块而非整会话终止)。
谢谢!
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
No repository files or tests are named. Start by reviewing the two request IDs and the reported multimodal and web_search scenarios, then determine whether the feedback can be addressed in this repository or requires service-side investigation. Done would require a clear response about the trigger location, applicable rule, and any supported diagnostic or fallback behavior.
Written by the indexing model from the issue text.
Assessment
- Domain
- ai, api, security
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100