Compare commits

...

7 Commits

Author SHA1 Message Date
dqy
21d1a6f3cc Merge branch 'main' into doc/add_banner
Some checks failed
Python application test / build (pull_request) Failing after 14m46s
2024-05-18 20:54:00 +08:00
dqy
dbdfc1897c Merge pull request 'doc:增加检测Python对其他语言的封装' (#17) from doc/python-binding into main
Some checks are pending
Python application test / build (push) Waiting to run
Reviewed-on: #17
Reviewed-by: sangge <sangge@noreply.localhost>
Reviewed-by: dqy <dqy@noreply.localhost>
2024-05-18 20:51:51 +08:00
594e0934e5 Merge branch 'main' into doc/python-binding
Some checks failed
Python application test / build (pull_request) Failing after 15m14s
2024-05-14 19:54:49 +08:00
252d9c655e Merge pull request 'fix: fix http proxy' (#19) from fix/fix-http-proxy into main
Some checks failed
Python application test / build (push) Failing after 12m54s
Reviewed-on: #19
2024-05-14 18:01:50 +08:00
80fff100b5 fix: fix http proxy
Some checks failed
Python application test / build (pull_request) Failing after 13m28s
2024-05-14 17:59:19 +08:00
3c7e5f4c3d doc:增加检测Python对其他语言的封装
Some checks are pending
Python application test / build (pull_request) Waiting to run
2024-05-11 21:37:35 +08:00
4bbe8ea62f doc:增加检测Python对其他语言的封装
Some checks failed
Python application test / build (pull_request) Failing after 12m6s
2024-05-05 20:12:43 +08:00
2 changed files with 12 additions and 1 deletions

View File

@@ -18,4 +18,4 @@ jobs:
run: python -m unittest discover -s tests
env:
OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
http_proxy: http://192.168.1.3:10809
http_proxy: http://192.168.1.2:10809

View File

@@ -8,6 +8,10 @@
参考项目: [https://github.com/SonarSource/sonarqube]
检查源代码的语法和关键词。通过这种方式,可以发现是否存在与其他语言的交互,比如调用外部命令、使用其他语言的扩展模块、与其他语言的接口交互等。
实现方法可以使用Python代码解析库如ast模块来分析语法树并检查特定的代码模式或结构开发脚本来搜索Python代码中常用于与其他语言交互的关键词和函数例如ctypes、subprocess、os.system等
## 控制流分析
通过分析程序的控制流(即程序中各个操作的执行顺序),可以检测到异常的控制流路径,这些路径可能是后门的迹象。
@@ -22,6 +26,10 @@
这个网站可以搜索依赖中是否存在漏洞: [https://security.snyk.io/package/pip/]
分析代码库中的依赖关系,查找是否导入了与其他语言交互相关的模块或库
实施策略:开发脚本进行依赖库对比匹配
## 异常行为检测
通过定义“正常”代码行为的基线,可以标识出异常行为,这些异常行为可能指示着后门的存在。
@@ -33,3 +41,6 @@
使用NLP技术来训练机器学习模型以自动从大量代码中学习和识别异常或潜在的后门模式。
开发方法采用深度学习框架如TensorFlow或PyTorch结合NLP处理工具训练模型识别代码中的异常行为。