Ovie
Ovie
我的博客是hexo 和 fluid 主题搭建的, fluid主题支持不蒜子,开启下就行,不用编辑代码。 因为不蒜子有移动端的问题,我转成了vercount,不过有些麻烦: 1. 修改 `_config.fluid.yml` (加上`?i=`以忽略fluid在后面添加的部分): ```yaml # busuanzi: https://busuanzi.ibruce.info/busuanzi/2.3/ busuanzi: https://cn.vercount.one/js?i= ``` 2. 修改 `themes/fluid/layout/_partial/scripts.ejs` , 加了段script: ```ejs function observeAndDisplay(sourceId, targetId) { if (document.getElementById(sourceId)) {...
在.env文件 里JWT Token硬编码。 https://github.com/huanghanzhilian/c-shopping/blob/1588741fe7631fd2712280dabce02253aeba5e99/.env#L3 以 http://shop.huanghanlian.com/ 为例,可以任意构造一个合法的JWT。  JWT里由userid组成,userid是MongoDB的ObjectID, Object ID可以预测,见 https://book.hacktricks.xyz/v/cn/network-services-pentesting/27017-27018-mongodb#mongo-objectid-yu-ce 。 修复建议: - 不使用硬编码的JWT密钥,项目初始化时随机生成
发送的http 1.0协议请求包错误: 
### Problem Description Some LLM response json string start with `\n` like: ``` {"choices":[{"message":{"content":"```\n{\"result\": 0}```","role":"assistant"},"index":0}]} ``` These cause parse LLM output error. (I test with Qwen2.5 and tool_calling_method is 'raw')...
I just checked then sources_data.json and the workflow log ( https://github.com/projectdiscovery/cdncheck/actions/runs/12530797903/job/34947671696 ) ,There are errors in the workflow log, resulting in some data being missing