TGTGInsighttelegram intelligenceLIVE / telegram public index
← GitHub Trends

TGINSIGHT SIMILAR POSTS

Find similar content

Source channel @githubtrending · Post #15510 · Feb 20

#go#ai_agents#ai_security_tool#anthropic#autonomous_agents#golang#gpt#graphql#multi_agent_system#offensive_security#open_source#openai#penetration_testing#penetration_testing_tools#react#security_automation#security_testing#security_tools#self_hosted PentAGI is an AI-powered tool that automates penetration testing with smart agents using 20+ pro tools like nmap and metasploit in a safe Docker sandbox. It researches vulnerabilities, executes attacks, stores knowledge for reuse, and creates detailed reports via a simple web UI. Quick setup needs Docker, an LLM API key (OpenAI/Anthropic), and `docker compose up -d`. This saves you hours of manual work, speeds up secure testing, cuts errors, and helps find issues faster for better protection. https://github.com/vxcontrol/pentagi

Results

1 similar post found

Search: #20b

当前筛选 #20b清除筛选
科技&趣闻&杂记

@kejiqu · Post #3730 · 08/27/2025, 02:50 AM

全球首个 AI 勒索软件现身:滥用 OpenAI 开源模型,跨平台搜索、窃取、加密文件 ESET 报告发现名为 PromptLock 的勒索软件,该软件是全球首个利用 AI 生成恶意代码的案例。PromptLock 调用 gpt-oss:20b 模型,在受感染设备本地生成恶意 Lua 代码,用于跨 Windows、Linux 和 macOS 系统搜索、窃取并加密文件。该模型可在具备 16GB 显存的设备上本地运行,支持自由修改和使用。攻击者可以通过建立内部智能体或隧道,将受害网络连接到外部服务器运行的模型,以规避对本地显存的需求。专家认为 PromptLock 或是概念验证程序,但预示着威胁行为者利用本地或私有 AI 的早期信号。OpenAI 表示已采取措施降低模型被恶意利用的风险。IT之家 🏷#PromptLock#AI#勒索软件#gpt#oss#20b 📢频道👥群组📝投稿