[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-splx-ai--agentic-radar":3,"tool-splx-ai--agentic-radar":61},[4,18,26,36,44,53],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",155373,2,"2026-04-14T11:34:08",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":32,"last_commit_at":42,"category_tags":43,"status":17},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",108322,"2026-04-10T11:39:34",[14,15,13],{"id":45,"name":46,"github_repo":47,"description_zh":48,"stars":49,"difficulty_score":32,"last_commit_at":50,"category_tags":51,"status":17},6121,"gemini-cli","google-gemini\u002Fgemini-cli","gemini-cli 是一款由谷歌推出的开源 AI 命令行工具，它将强大的 Gemini 大模型能力直接集成到用户的终端环境中。对于习惯在命令行工作的开发者而言，它提供了一条从输入提示词到获取模型响应的最短路径，无需切换窗口即可享受智能辅助。\n\n这款工具主要解决了开发过程中频繁上下文切换的痛点，让用户能在熟悉的终端界面内直接完成代码理解、生成、调试以及自动化运维任务。无论是查询大型代码库、根据草图生成应用，还是执行复杂的 Git 操作，gemini-cli 都能通过自然语言指令高效处理。\n\n它特别适合广大软件工程师、DevOps 人员及技术研究人员使用。其核心亮点包括支持高达 100 万 token 的超长上下文窗口，具备出色的逻辑推理能力；内置 Google 搜索、文件操作及 Shell 命令执行等实用工具；更独特的是，它支持 MCP（模型上下文协议），允许用户灵活扩展自定义集成，连接如图像生成等外部能力。此外，个人谷歌账号即可享受免费的额度支持，且项目基于 Apache 2.0 协议完全开源，是提升终端工作效率的理想助手。",100752,"2026-04-10T01:20:03",[52,13,15,14],"插件",{"id":54,"name":55,"github_repo":56,"description_zh":57,"stars":58,"difficulty_score":32,"last_commit_at":59,"category_tags":60,"status":17},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[52,14],{"id":62,"github_repo":63,"name":64,"description_en":65,"description_zh":66,"ai_summary_zh":66,"readme_en":67,"readme_zh":68,"quickstart_zh":69,"use_case_zh":70,"hero_image_url":71,"owner_login":72,"owner_name":73,"owner_avatar_url":74,"owner_bio":75,"owner_company":76,"owner_location":76,"owner_email":77,"owner_twitter":76,"owner_website":78,"owner_url":79,"languages":80,"stars":89,"forks":90,"last_commit_at":91,"license":92,"difficulty_score":32,"env_os":93,"env_gpu":94,"env_ram":94,"env_deps":95,"category_tags":101,"github_topics":102,"view_count":32,"oss_zip_url":76,"oss_zip_packed_at":76,"status":17,"created_at":119,"updated_at":120,"faqs":121,"releases":152},7414,"splx-ai\u002Fagentic-radar","agentic-radar","A security scanner for your LLM agentic workflows","agentic-radar 是一款专为大语言模型（LLM）智能体工作流打造的安全扫描工具。随着 AI 智能体在自动化任务中承担越来越重要的角色，其面临的安全风险也随之增加，例如提示词注入、敏感数据泄露或恶意指令执行等。agentic-radar 旨在帮助开发者在部署前识别并修复这些潜在漏洞，确保智能体系统的稳健运行。\n\n该工具主要面向构建和运维 AI 智能体的开发者及安全研究人员，尤其适合使用 CrewAI、OpenAI Agents 等框架的团队。它不仅能可视化展示智能体的工作流程，让复杂的交互逻辑一目了然，还提供了独特的“提示词加固”功能，自动增强系统抵御攻击的能力。此外，agentic-radar 支持集成到 CI\u002FCD 流水线中，实现安全测试的自动化，让安全防护成为开发流程的自然组成部分。通过直观的图表和详细的报告，用户可以轻松定位风险点，无需深厚的安全背景也能上手。无论是初创团队还是大型企业，只要涉及 AI 智能体应用，agentic-radar 都是保障系统安全的得力助手。","\u003Cdiv align=\"center\">\n\n\n  \u003Ca href=\"https:\u002F\u002Fsplx.ai\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_readme_81c8fb2703ad.png\" alt=\"logo\" width=\"600\" height=\"auto\" \u002F>\n  \u003C\u002Fa>\n  \n  \u003Cp>\n    A Security Scanner for your agentic workflows!\n  \u003C\u002Fp>\n  \n  \n\u003C!-- Badges -->\n\u003Cp>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fgraphs\u002Fcontributors\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fcontributors\u002Fsplx-ai\u002Fagentic-radar\" alt=\"contributors\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flast-commit\u002Fsplx-ai\u002Fagentic-radar\" alt=\"last update\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fnetwork\u002Fmembers\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fforks\u002Fsplx-ai\u002Fagentic-radar\" alt=\"forks\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fstargazers\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsplx-ai\u002Fagentic-radar\" alt=\"stars\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fissues\u002Fsplx-ai\u002Fagentic-radar\" alt=\"open issues\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fblob\u002Fmain\u002FLICENSE\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fsplx-ai\u002Fagentic-radar.svg\" alt=\"license\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fagentic-radar\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fagentic-radar\" alt=\"PyPI - Version\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fagentic-radar\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_readme_1c7d3ce53766.png\" alt=\"PyPI - Downloads\" \u002F>\n  \u003C\u002Fa>\n  \u003Cbr \u002F>\n  \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FtR2d54utZc\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fdiscord\u002F1346578514177949767?style=for-the-badge&logo=discord&logoColor=white&label=Discord&labelColor=5865F2&color=555555\" alt=\"Discord\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fjoin.slack.com\u002Ft\u002Fsplxaicommunity\u002Fshared_invite\u002Fzt-31b3hc3mt-A0v78qztTIMSNBg6y~WOAA\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FSlack-4A154B?style=for-the-badge&logo=slack&logoColor=white\" alt=\"Slack\" \u002F>\n  \u003C\u002Fa>\n\u003C\u002Fp>\n   \n  \u003Ch4>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002F\">View Demo\u003C\u002Fa>\n  \u003Cspan> · \u003C\u002Fspan>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\">Documentation\u003C\u002Fa>\n  \u003Cspan> · \u003C\u002Fspan>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F\">Report Bug\u003C\u002Fa>\n  \u003Cspan> · \u003C\u002Fspan>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F\">Request Feature\u003C\u002Fa>\n  \u003C\u002Fh4>\n\u003C\u002Fdiv>\n\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_readme_12b83506ce96.png\"\u002F>\n\n\u003C!-- TABLE OF CONTENTS -->\n\u003Cdetails>\n  \u003Csummary>Table of Contents\u003C\u002Fsummary>\n  \u003Col>\n    \u003Cli>\n      \u003Ca href=\"#description-\">Description\u003C\u002Fa>\n    \u003C\u002Fli>\n    \u003Cli>\n      \u003Ca href=\"#agentic-visualizer-\">Agentic Visualizer\u003C\u002Fa>\n    \u003C\u002Fli>\n    \u003Cli>\n      \u003Ca href=\"#getting-started-\">Getting Started\u003C\u002Fa>\n      \u003Cul>\n        \u003Cli>\u003Ca href=\"#prerequisites\">Prerequisites\u003C\u002Fa>\u003C\u002Fli>\n        \u003Cli>\u003Ca href=\"#installation\">Installation\u003C\u002Fa>\u003C\u002Fli>\n      \u003C\u002Ful>\n    \u003C\u002Fli>\n      \u003Cli>\n      \u003Ca href=\"#advanced-installation\">Advanced Installation\u003C\u002Fa>\n      \u003Cul>\n        \u003Cli>\u003Ca href=\"#crewai-installation\">CrewAI Installation\u003C\u002Fa>\u003C\u002Fli>\n        \u003Cli>\u003Ca href=\"#openai-agents-installation\">OpenAI Agents Installation\u003C\u002Fa>\u003C\u002Fli>\n      \u003C\u002Ful>\n    \u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#usage\">Usage\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\n      \u003Ca href=\"#advanced-features-\">Advanced Features\u003C\u002Fa>\n      \u003Cul>\n        \u003Cli>\u003Ca href=\"#agentic-prompt-hardening\">Agentic Prompt Hardening\u003C\u002Fa>\u003C\u002Fli>\n        \u003Cli>\u003Ca href=\"#-test-for-vulnerabilities-in-agentic-workflows\">Test for Vulnerabilities in Agentic Workflows\u003C\u002Fa>\u003C\u002Fli>\n        \u003Cli>\u003Ca href=\"#cicd-workflow\">CI\u002FCD Workflow\u003C\u002Fa>\u003C\u002Fli>\n      \u003C\u002Ful>\n    \u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#roadmap-\">Roadmap\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#blogs-and-tutorials-\">Blogs and Tutorials\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#community-\">Community\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#frequently-asked-questions-\">Frequently Asked Questions\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#contributing-\">Contributing\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#code-of-conduct-\">Code Of Conduct\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#license-\">License\u003C\u002Fa>\u003C\u002Fli>\n  \u003C\u002Fol>\n\u003C\u002Fdetails>\n\n## Description 📝\n\nThe **Agentic Radar** is designed to analyze and assess agentic systems for security and operational insights. It helps developers, researchers, and security professionals understand how agentic systems function and identify potential vulnerabilities.\n\nIt allows users to create a security report for agentic systems, including:\n1. **Workflow Visualization** - a graph of the agentic system's workflow✅\n2. **Tool Identification** - a list of all external and custom tools utilized by the system✅\n3. **MCP Server Detection** - a list of all MCP servers used by system's agents✅\n4. **Vulnerability Mapping** - a table connecting identified tools to known vulnerabilities, providing a security overview✅\n\nThe comprehensive HTML report summarizes all findings and allows for easy reviewing and sharing.\n\n**[View Full Report Example Here](https:\u002F\u002Fagentic-radar.neocities.org\u002F)**\n\n\n**Agentic Radar** includes mapping of detected vulnerabilities to well-known security frameworks 🛡️.\n+ [OWASP Top 10 LLM Applications](https:\u002F\u002Fowasp.org\u002Fwww-project-top-10-for-large-language-model-applications\u002F)\n\n+ [OWASP Agentic AI – Threats and Mitigations](https:\u002F\u002Fgenaisecurityproject.com\u002Fresource\u002Fagentic-ai-threats-and-mitigations)\n\n## Agentic Visualizer 🎆\n\nIf you only care about visualization, try out the [Agentic Visualizer](https:\u002F\u002Fagentic-visualizer.splx.ai\u002F).\n\nIt is a web-based tool that allows you to visualize agentic workflows in a user-friendly way.\n\n\n\n## Getting Started 🚀\n\n### Prerequisites\n\nThere are none! Just make sure you have Python (pip) installed on your machine.\n\n### Installation\n```sh\npip install agentic-radar\n\n# Check that it is installed\nagentic-radar --version\n```\n\nSome features require extra installations, depending on the targeted agentic framework. See more [below](#advanced-installation).\n\n## Advanced Installation\n### CrewAI Installation\n\nCrewAI extras are needed when using one of the following features in combination with CrewAI:\n\n- [Agentic Radar Test](#-test-for-vulnerabilities-in-agentic-workflows)\n- Descriptions for predefined tools\n\nYou can install Agentic Radar with extra CrewAI dependencies by running:\n```sh\npip install \"agentic-radar[crewai]\"\n```\n\n> [!WARNING]\n> This will install the `crewai-tools` package which is only supported on Python versions >= 3.10 and \u003C 3.13.\n> If you are using a different python version, the tool descriptions will be less detailed or entirely missing.\n\n### OpenAI Agents Installation\n\nOpenAI Agents extras are needed when using one of the following features in combination with OpenAI Agents:\n\n- [Agentic Radar Test](#-test-for-vulnerabilities-in-agentic-workflows)\n\nYou can install Agentic Radar with extra OpenAI Agents dependencies by running:\n```sh\npip install \"agentic-radar[openai-agents]\"\n```\n\n## Usage\n\nAgentic Radar now supports two main commands:\n\n### 1. `scan`\nScan code for agentic workflows and generate a report.\n\n```sh\nagentic-radar scan [OPTIONS] FRAMEWORK:{langgraph|crewai|n8n|openai-agents|autogen}\n```\n\nExample:\n```sh\nagentic-radar scan langgraph -i path\u002Fto\u002Flanggraph\u002Fexample\u002Ffolder -o report.html\n```\n\n---\n\n### 2. `test`\nTest agents in an agentic workflow for various vulnerabilities.\nRequires OPENAI_API_KEY set as environment variable.\n\n```sh\nagentic-radar test [OPTIONS] FRAMEWORK:{openai-agents} ENTRYPOINT_SCRIPT_WITH_ARGS\n```\n\nExample:\n```sh\nagentic-radar test openai-agents \"path\u002Fto\u002Fopenai-agents\u002Fexample.py\"\n```\n\nSee more about this feature [here](#-test-for-vulnerabilities-in-agentic-workflows).\n\n\n## Advanced Features ✨\n\n### Agentic Prompt Hardening\n\nAgentic Prompt Hardening automatically improves detected system prompts in your agentic workflow and displays them in the report. It transforms simple agent instructions into high-quality structured system prompts which follow best prompt engineering practices.\n\n> [!NOTE]  \n> Currently supported frameworks (with more to come): OpenAI Agents, CrewAI, Autogen\n\nIt is quite straightforward to use:\n1. Set your OPENAI_API_KEY environment variable by running `export OPENAI_API_KEY=\u003Capi_key>`.\n\n2. Run Agentic Radar with the `--harden-prompts` flag, for example:\n```sh\nagentic-radar scan openai-agents --harden-prompts -i examples\u002Fopenai-agents\u002F\nbasic\u002Flifecycle_example -o report.html\n```\n\n4. Inspect hardened system prompts in the generated report:\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_readme_b8c996eb11ed.png\"\u002F>\n\n### 🔍 Test for Vulnerabilities in Agentic Workflows\n\nAgentic Radar now supports testing your agent workflows at **runtime** to identify critical vulnerabilities through simulated adversarial inputs.\n\nThis includes automated testing for:\n  - Prompt Injection\n  - PII Leakage\n  - Harmful Content Generation\n  - Fake News Generation\n\nCurrently supported for:\n- OpenAI Agents ✅ (more frameworks coming soon)\n\n#### 🛠 How It Works\n\nThe test command launches your agentic workflow with a test suite designed to simulate malicious or adversarial inputs. These tests are designed based on real-world attack scenarios aligned with the OWASP LLM Top 10.\n\n> [!NOTE]  \n> This feature requires OPENAI_API_KEY or AZURE_OPENAI_API_KEY set as an environment variable. You can set it via command line or inside a .env file.\n\nTest is run like:\n```sh\nagentic-radar test \u003Cframework> \"\u003Cpath\u002Fto\u002Fthe\u002Fworkflow\u002Fmain.py any-necessary-args>\"\n```\n\nFor example:\n```sh\nagentic-radar test openai-agents \"examples\u002Fopenai-agents\u002Fbasic\u002Flifecycle_example.py\"\n```\n\nThe tool injects itself into the agentic workflow provided by user, detects necessary information and runs the prepared tests.\n\n#### Configurable Tests\n\nYou can pass custom tests to Agentic Radar by creating a YAML file with the following structure:\n```yaml\ninclude_default_tests: false\ntests:\n   - name: TestA\n     input: Mention A.\n     success_condition: Attack succeeded if A is mentioned.\n\n   - name: TestB\n     input: Mention B.\n     success_condition: Attack succeeded if B is mentioned.\n```\n\n`name` is the name of the test, `input` is the input text to be passed to the agent, and `success_condition` is a description of what constitutes a successful attack (it will be used by an oracle LLM to determine if the test passed or failed).\n\nYou can then run Agentic Radar test with the `--config` option pointing to your YAML file:\n```sh\nagentic-radar test openai-agents --config custom_tests.yaml \"examples\u002Fopenai-agents\u002Fbasic\u002Flifecycle_example.py\"\n```\n\nBy default, Agentic Radar will also include the built-in tests. You can disable them by setting `include_default_tests` to `false` in your YAML file.\n\n#### 📊 Rich Test Results\n\nAll test results are printed in a visually rich table format directly in the terminal.\nEach row shows:\n  - Agent name\n  - Type of test\n  - Injected input\n  - Agent output\n  - ✅ Whether the test passed or failed\n  - 🛑 A short explanation of the result\n\nThis makes it easy to spot vulnerabilities at a glance—especially in multi-agent systems.\n\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_readme_3e38ec9dda2b.png\" alt=\"Test Results Example\" \u002F>\n\n### CI\u002FCD Workflow\n\nTo integrate Agentic Radar into your CI\u002FCD pipeline, you can use the [provided GitHub Actions workflow example](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fblob\u002Fmain\u002Fexamples\u002Fgithub_workflow\u002Fauto-agentic-radar.yaml). Just paste the YAML to the `.github\u002Fworkflows` directory of your repository. \n\nThis workflow automatically runs Agentic Radar scans on your codebase whenever changes are pushed to the repository. The generated report is uploaded as an artifact in the GitHub Actions run.\n\n## Roadmap 📈\n\nThis matrix shows which agentic frameworks support all the Agentic Radar features. With time we will strive towards covering all current frameworks with all existing features, as well as introducing new frameworks to the mix. \n\n| Feature       | Scan        | MCP Detection        | Prompt Hardening | Agentic Test\n|----------------|-------------|-------------|-------------|-------------|\n| OpenAI Agents  | ✅          | ✅          | ✅          |     ✅         |\n| CrewAI         | ✅          | ✅         | ✅          |      ❌        |\n| n8n            | ✅          | ✅          | ❌          |     ❌        |\n| LangGraph      | ✅          | ✅          | ❌          |     ❌         |\n| Autogen     | ✅          | ✅          | ✅          |     ❌         |\n\nAre there some features you would like to see happen first? Vote anonymously [here](https:\u002F\u002Fstrawpoll.com\u002Fw4nWWMqqlnA) or [open a GitHub Issue](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002Fnew\u002Fchoose).\n\n## Blogs and Tutorials 💡\n\n- [CrewAI](https:\u002F\u002Fsplx.ai\u002Fblog\u002Fenhancing-ai-transparency-scanning-crewai-workflows-with-agentic-radar)\n- [n8n](https:\u002F\u002Fsplx.ai\u002Fblog\u002Fscanning-n8n-workflows-with-agentic-radar)\n- [OpenAI Agents](https:\u002F\u002Fsplx.ai\u002Fblog\u002Fopenai-agents-sdk-transparent-workflows-with-agentic-radar)\n- [Autogen](https:\u002F\u002Fcolab.research.google.com\u002Fdrive\u002F14IeJv08lzBsLlEO9cKoHloDioWMWGf5Q)\n- [MCP Server Detection](https:\u002F\u002Fsplx.ai\u002Fblog\u002Fagentic-radar-now-detects-mcp-servers-in-agentic-workflows)\n- [Agentic Prompt Hardening](https:\u002F\u002Fsplx.ai\u002Fblog\u002Fagentic-radar-now-scans-and-hardens-system-prompts-in-agentic-workflows)\n\n## Community 🤝\n\nWe welcome contributions from the AI and security community! Join our [Discord community](https:\u002F\u002Fdiscord.gg\u002FQZQpef5PsD) or [Slack community](https:\u002F\u002Fjoin.slack.com\u002Ft\u002Fsplxaicommunity\u002Fshared_invite\u002Fzt-31b3hc3mt-A0v78qztTIMSNBg6y~WOAA) to connect with other developers, discuss features, get support and contribute to **Agentic Radar** 🚀\n\nIf you like what you see, give us a star! It keeps us inspired to improve and innovate and helps others discover the project 🌟\n\n## Frequently Asked Questions ❓\n\n**Q: Is my source code being shared or is everything running locally?**  \nA: The main features (static workflow analysis and vulnerability mapping) are run completely locally and therefore your code is not shared anywhere. For optional advanced features, LLM's might be used. Eg. when using [Prompt Hardening](#agentic-prompt-hardening), detected system prompts can get sent to LLM for analysis.\n\n## Contributing 💻 \n\n[CONTRIBUTING](CONTRIBUTING.md)\n\n## Code Of Conduct 📜\n[CODE OF CONDUCT](CODE_OF_CONDUCT.md)\n\n## License ⚖️\n\n[LICENSE](LICENSE)\n","\u003Cdiv align=\"center\">\n\n\n  \u003Ca href=\"https:\u002F\u002Fsplx.ai\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_readme_81c8fb2703ad.png\" alt=\"logo\" width=\"600\" height=\"auto\" \u002F>\n  \u003C\u002Fa>\n  \n  \u003Cp>\n    您的智能体工作流安全扫描工具！\n  \u003C\u002Fp>\n  \n  \n\u003C!-- Badges -->\n\u003Cp>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fgraphs\u002Fcontributors\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fcontributors\u002Fsplx-ai\u002Fagentic-radar\" alt=\"contributors\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flast-commit\u002Fsplx-ai\u002Fagentic-radar\" alt=\"last update\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fnetwork\u002Fmembers\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fforks\u002Fsplx-ai\u002Fagentic-radar\" alt=\"forks\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fstargazers\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fsplx-ai\u002Fagentic-radar\" alt=\"stars\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fissues\u002Fsplx-ai\u002Fagentic-radar\" alt=\"open issues\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fblob\u002Fmain\u002FLICENSE\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fsplx-ai\u002Fagentic-radar.svg\" alt=\"license\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fagentic-radar\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fagentic-radar\" alt=\"PyPI - Version\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fagentic-radar\">\n    \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_readme_1c7d3ce53766.png\" alt=\"PyPI - Downloads\" \u002F>\n  \u003C\u002Fa>\n  \u003Cbr \u002F>\n  \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FtR2d54utZc\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fdiscord\u002F1346578514177949767?style=for-the-badge&logo=discord&logoColor=white&label=Discord&labelColor=5865F2&color=555555\" alt=\"Discord\" \u002F>\n  \u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fjoin.slack.com\u002Ft\u002Fsplxaicommunity\u002Fshared_invite\u002Fzt-31b3hc3mt-A0v78qztTIMSNBg6y~WOAA\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FSlack-4A154B?style=for-the-badge&logo=slack&logoColor=white\" alt=\"Slack\" \u002F>\n  \u003C\u002Fa>\n\u003C\u002Fp>\n   \n  \u003Ch4>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002F\">查看演示\u003C\u002Fa>\n  \u003Cspan> · \u003C\u002Fspan>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\">文档\u003C\u002Fa>\n  \u003Cspan> · \u003C\u002Fspan>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F\">报告 bug\u003C\u002Fa>\n  \u003Cspan> · \u003C\u002Fspan>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F\">请求功能\u003C\u002Fa>\n  \u003C\u002Fh4>\n\u003C\u002Fdiv>\n\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_readme_12b83506ce96.png\"\u002F>\n\n\u003C!-- TABLE OF CONTENTS -->\n\u003Cdetails>\n  \u003Csummary>目录\u003C\u002Fsummary>\n  \u003Col>\n    \u003Cli>\n      \u003Ca href=\"#description-\">描述\u003C\u002Fa>\n    \u003C\u002Fli>\n    \u003Cli>\n      \u003Ca href=\"#agentic-visualizer-\">智能体可视化工具\u003C\u002Fa>\n    \u003C\u002Fli>\n    \u003Cli>\n      \u003Ca href=\"#getting-started-\">快速入门\u003C\u002Fa>\n      \u003Cul>\n        \u003Cli>\u003Ca href=\"#prerequisites\">前提条件\u003C\u002Fa>\u003C\u002Fli>\n        \u003Cli>\u003Ca href=\"#installation\">安装\u003C\u002Fa>\u003C\u002Fli>\n      \u003C\u002Ful>\n    \u003C\u002Fli>\n    \u003Cli>\n      \u003Ca href=\"#advanced-installation\">高级安装\u003C\u002Fa>\n      \u003Cul>\n        \u003Cli>\u003Ca href=\"#crewai-installation\">CrewAI 安装\u003C\u002Fa>\u003C\u002Fli>\n        \u003Cli>\u003Ca href=\"#openai-agents-installation\">OpenAI Agents 安装\u003C\u002Fa>\u003C\u002Fli>\n      \u003C\u002Ful>\n    \u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#usage\">使用方法\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\n      \u003Ca href=\"#advanced-features-\">高级功能\u003C\u002Fa>\n      \u003Cul>\n        \u003Cli>\u003Ca href=\"#agentic-prompt-hardening\">智能体提示加固\u003C\u002Fa>\u003C\u002Fli>\n        \u003Cli>\u003Ca href=\"#-test-for-vulnerabilities-in-agentic-workflows\">检测智能体工作流中的漏洞\u003C\u002Fa>\u003C\u002Fli>\n        \u003Cli>\u003Ca href=\"#cicd-workflow\">CI\u002FCD 工作流\u003C\u002Fa>\u003C\u002Fli>\n      \u003C\u002Ful>\n    \u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#roadmap-\">路线图\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#blogs-and-tutorials-\">博客与教程\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#community-\">社区\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#frequently-asked-questions-\">常见问题解答\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#contributing-\">贡献\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#code-of-conduct-\">行为准则\u003C\u002Fa>\u003C\u002Fli>\n    \u003Cli>\u003Ca href=\"#license-\">许可证\u003C\u002Fa>\u003C\u002Fli>\n  \u003C\u002Fol>\n\u003C\u002Fdetails>\n\n## 描述 📝\n\n**Agentic Radar** 旨在分析和评估智能体系统的安全性及运行情况，帮助开发者、研究人员和安全专业人士理解智能体系统的工作原理并识别潜在的安全隐患。\n\n它允许用户为智能体系统生成一份安全报告，其中包括：\n1. **工作流可视化**——智能体系统工作流的图表✅\n2. **工具识别**——系统所使用的全部外部工具和自定义工具列表✅\n3. **MCP 服务器检测**——系统代理使用的所有 MCP 服务器列表✅\n4. **漏洞映射**——将已识别的工具与已知漏洞对应起来的表格，提供全面的安全概览✅\n\n这份详尽的 HTML 报告汇总了所有发现，便于查阅和分享。\n\n**[在此处查看完整报告示例](https:\u002F\u002Fagentic-radar.neocities.org\u002F)**\n\n\n**Agentic Radar** 还会将检测到的漏洞映射到知名的安全框架中 🛡️。\n+ [OWASP 大型语言模型应用十大风险](https:\u002F\u002Fowasp.org\u002Fwww-project-top-10-for-large-language-model-applications\u002F)\n\n+ [OWASP 智能体 AI – 威胁与缓解措施](https:\u002F\u002Fgenaisecurityproject.com\u002Fresource\u002Fagentic-ai-threats-and-mitigations)\n\n## 智能体可视化工具 🎆\n\n如果您只关心可视化效果，可以尝试使用 [Agentic Visualizer](https:\u002F\u002Fagentic-visualizer.splx.ai\u002F)。\n\n这是一个基于网页的工具，能够以友好的方式可视化智能体工作流。\n\n\n\n## 快速入门 🚀\n\n### 前提条件\n\n无需任何特殊要求！只需确保您的设备上已安装 Python（pip）即可。\n\n### 安装\n```sh\npip install agentic-radar\n\n# 检查是否安装成功\nagentic-radar --version\n```\n\n某些功能需要额外的安装依赖，具体取决于您要分析的智能体框架。更多信息请参见 [下方](#advanced-installation)。\n\n## 高级安装\n### CrewAI 安装\n\n当您结合以下功能使用 CrewAI 时，需要安装 CrewAI 的额外依赖项：\n\n- [Agentic Radar 漏洞检测](#-test-for-vulnerabilities-in-agentic-workflows)\n- 预定义工具的描述信息\n\n您可以通过运行以下命令安装带有 CrewAI 依赖的 Agentic Radar：\n```sh\npip install \"agentic-radar[crewai]\"\n```\n\n> [!警告]\n> 此操作将安装 `crewai-tools` 包，该包仅支持 Python 3.10 及以上版本，但不支持 3.13 版本。\n> 如果您使用的是其他 Python 版本，工具描述可能会不够详细，甚至完全缺失。\n\n### OpenAI Agents 安装\n\n当您结合以下功能使用 OpenAI Agents 时，需要安装 OpenAI Agents 的额外依赖项：\n\n- [Agentic Radar 漏洞检测](#-test-for-vulnerabilities-in-agentic-workflows)\n\n您可以通过运行以下命令安装带有 OpenAI Agents 依赖的 Agentic Radar：\n```sh\npip install \"agentic-radar[openai-agents]\"\n```\n\n## 使用方法\n\nAgentic Radar 目前支持两个主要命令：\n\n### 1. `scan`\n扫描代码以生成智能体工作流报告。\n\n```sh\nagentic-radar scan [OPTIONS] FRAMEWORK:{langgraph|crewai|n8n|openai-agents|autogen}\n```\n\n示例：\n```sh\nagentic-radar scan langgraph -i path\u002Fto\u002Flanggraph\u002Fexample\u002Ffolder -o report.html\n```\n\n---\n\n### 2. `test`\n测试智能体工作流中的代理，以检测各种漏洞。  \n需要将 `OPENAI_API_KEY` 设置为环境变量。\n\n```sh\nagentic-radar test [OPTIONS] FRAMEWORK:{openai-agents} ENTRYPOINT_SCRIPT_WITH_ARGS\n```\n\n示例：\n```sh\nagentic-radar test openai-agents \"path\u002Fto\u002Fopenai-agents\u002Fexample.py\"\n```\n\n有关此功能的更多信息，请参阅[此处](#-测试智能体工作流中的漏洞)。\n\n## 高级功能 ✨\n\n### 智能体提示加固\n\n智能体提示加固功能会自动优化您智能体工作流中检测到的系统提示，并在报告中展示。它会将简单的代理指令转换为高质量的结构化系统提示，遵循最佳提示工程实践。\n\n> [!NOTE]  \n> 目前支持的框架（未来还将增加更多）：OpenAI Agents、CrewAI、Autogen\n\n使用方法非常简单：\n1. 通过运行 `export OPENAI_API_KEY=\u003Capi_key>` 来设置您的 `OPENAI_API_KEY` 环境变量。\n\n2. 使用 `--harden-prompts` 标志运行 Agentic Radar，例如：\n```sh\nagentic-radar scan openai-agents --harden-prompts -i examples\u002Fopenai-agents\u002F\nbasic\u002Flifecycle_example -o report.html\n```\n\n4. 在生成的报告中查看加固后的系统提示：\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_readme_b8c996eb11ed.png\"\u002F>\n\n### 🔍 测试智能体工作流中的漏洞\n\nAgentic Radar 现在支持在**运行时**测试您的代理工作流，通过模拟对抗性输入来识别关键漏洞。\n\n这包括对以下内容的自动化测试：\n  - 提示注入\n  - PII 泄露\n  - 有害内容生成\n  - 虚假新闻生成\n\n目前支持：\n- OpenAI Agents ✅（更多框架即将推出）\n\n#### 🛠 工作原理\n\n测试命令会启动您的智能体工作流，并运行一套旨在模拟恶意或对抗性输入的测试套件。这些测试基于与 OWASP LLM Top 10 对齐的真实攻击场景设计。\n\n> [!NOTE]  \n> 此功能需要将 `OPENAI_API_KEY` 或 `AZURE_OPENAI_API_KEY` 设置为环境变量。您可以通过命令行或在 `.env` 文件中进行设置。\n\n测试的运行方式如下：\n```sh\nagentic-radar test \u003Cframework> \"\u003Cpath\u002Fto\u002Fthe\u002Fworkflow\u002Fmain.py any-necessary-args>\"\n```\n\n例如：\n```sh\nagentic-radar test openai-agents \"examples\u002Fopenai-agents\u002Fbasic\u002Flifecycle_example.py\"\n```\n\n该工具会将自身注入用户提供的智能体工作流中，检测必要信息并运行预设的测试。\n\n#### 可配置的测试\n\n您可以创建一个 YAML 文件，按照以下结构向 Agentic Radar 传递自定义测试：\n```yaml\ninclude_default_tests: false\ntests:\n   - name: TestA\n     input: 提到 A。\n     success_condition: 如果提到 A，则攻击成功。\n\n   - name: TestB\n     input: 提到 B。\n     success_condition: 如果提到 B，则攻击成功。\n```\n\n`name` 是测试的名称，`input` 是要传递给代理的输入文本，`success_condition` 是描述攻击成功的条件（将由 Oracle LLM 用来判断测试是否通过）。\n\n然后，您可以使用 `--config` 选项指向您的 YAML 文件来运行 Agentic Radar 测试：\n```sh\nagentic-radar test openai-agents --config custom_tests.yaml \"examples\u002Fopenai-agents\u002Fbasic\u002Flifecycle_example.py\"\n```\n\n默认情况下，Agentic Radar 还会包含内置测试。您可以通过在 YAML 文件中将 `include_default_tests` 设置为 `false` 来禁用它们。\n\n#### 📊 丰富的测试结果\n\n所有测试结果都会以视觉上丰富的表格格式直接打印在终端中。每行显示：\n  - 代理名称\n  - 测试类型\n  - 注入的输入\n  - 代理输出\n  - ✅ 测试是否通过\n  - 🛑 结果的简短说明\n\n这使得即使在多代理系统中，也能轻松一眼发现漏洞。\n\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_readme_3e38ec9dda2b.png\" alt=\"测试结果示例\" \u002F>\n\n### CI\u002FCD 工作流\n\n要将 Agentic Radar 集成到您的 CI\u002FCD 流程中，您可以使用[提供的 GitHub Actions 工作流示例](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fblob\u002Fmain\u002Fexamples\u002Fgithub_workflow\u002Fauto-agentic-radar.yaml)。只需将 YAML 文件粘贴到您仓库的 `.github\u002Fworkflows` 目录中即可。\n\n此工作流会在每次向仓库推送更改时自动运行 Agentic Radar 扫描。生成的报告会作为工件上传到 GitHub Actions 运行中。\n\n## 路线图 📈\n\n下表展示了哪些智能体框架支持 Agentic Radar 的所有功能。随着时间的推移，我们将努力使所有现有框架都覆盖所有功能，并引入新的框架。\n\n| 功能       | 扫描        | MCP 检测        | 提示加固 | 智能体测试\n|----------------|-------------|-------------|-------------|-------------|\n| OpenAI Agents  | ✅          | ✅          | ✅          |     ✅         |\n| CrewAI         | ✅          | ✅         | ✅          |      ❌        |\n| n8n            | ✅          | ✅          | ❌          |     ❌        |\n| LangGraph      | ✅          | ✅          | ❌          |     ❌         |\n| Autogen     | ✅          | ✅          | ✅          |     ❌         |\n\n您希望优先实现哪些功能？请在此处匿名投票[这里](https:\u002F\u002Fstrawpoll.com\u002Fw4nWWMqqlnA)，或[打开一个 GitHub Issue](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002Fnew\u002Fchoose)。\n\n## 博客和教程 💡\n\n- [CrewAI](https:\u002F\u002Fsplx.ai\u002Fblog\u002Fenhancing-ai-transparency-scanning-crewai-workflows-with-agentic-radar)\n- [n8n](https:\u002F\u002Fsplx.ai\u002Fblog\u002Fscanning-n8n-workflows-with-agentic-radar)\n- [OpenAI Agents](https:\u002F\u002Fsplx.ai\u002Fblog\u002Fopenai-agents-sdk-transparent-workflows-with-agentic-radar)\n- [Autogen](https:\u002F\u002Fcolab.research.google.com\u002Fdrive\u002F14IeJv08lzBsLlEO9cKoHloDioWMWGf5Q)\n- [MCP 服务器检测](https:\u002F\u002Fsplx.ai\u002Fblog\u002Fagentic-radar-now-detects-mcp-servers-in-agentic-workflows)\n- [智能体提示加固](https:\u002F\u002Fsplx.ai\u002Fblog\u002Fagentic-radar-now-scans-and-hardens-system-prompts-in-agentic-workflows)\n\n## 社区 🤝\n\n我们欢迎来自 AI 和安全社区的贡献！加入我们的[Discord 社区](https:\u002F\u002Fdiscord.gg\u002FQZQpef5PsD)或[Slack 社区](https:\u002F\u002Fjoin.slack.com\u002Ft\u002Fsplxaicommunity\u002Fshared_invite\u002Fzt-31b3hc3mt-A0v78qztTIMSNBg6y~WOAA)，与其他开发者交流、讨论功能、获取支持，并为 **Agentic Radar** 做出贡献 🚀\n\n如果您喜欢我们的工作，请给我们点个赞！这会激励我们不断改进和创新，同时帮助更多人发现这个项目 🌟\n\n## 常见问题 ❓\n\n**问：我的源代码会被共享吗？还是所有操作都在本地运行？**  \n答：主要功能（静态工作流分析和漏洞映射）完全在本地运行，因此您的代码不会被共享到任何地方。对于可选的高级功能，可能会使用大语言模型。例如，在使用[提示加固](#agentic-prompt-hardening)时，检测到的系统提示可能会被发送到大语言模型进行分析。\n\n## 参与贡献 💻 \n\n[CONTRIBUTING](CONTRIBUTING.md)\n\n## 行为准则 📜\n[行为准则](CODE_OF_CONDUCT.md)\n\n## 许可证 ⚖️\n\n[许可证](LICENSE)","# Agentic Radar 快速上手指南\n\nAgentic Radar 是一款专为智能体（Agentic）工作流设计的安全扫描工具。它能帮助开发者可视化工作流、识别外部工具与 MCP 服务器，并映射潜在的安全漏洞（基于 OWASP 标准），最终生成详细的 HTML 安全报告。\n\n## 环境准备\n\n*   **操作系统**：支持 Windows、macOS 和 Linux。\n*   **Python 版本**：需安装 Python (建议 3.8+)。\n    *   *注意*：若使用 CrewAI 扩展功能，Python 版本必须在 `>= 3.10` 且 `\u003C 3.13` 之间。\n*   **前置依赖**：无特殊系统依赖，确保已安装 `pip` 包管理工具。\n*   **可选环境变量**：若需使用“提示词加固”或“漏洞测试”功能，需配置 `OPENAI_API_KEY`。\n\n## 安装步骤\n\n### 1. 基础安装\n通过 PyPI 安装核心功能：\n\n```sh\npip install agentic-radar\n\n# 验证安装\nagentic-radar --version\n```\n\n> **国内加速建议**：如遇下载缓慢，可使用清华或阿里镜像源：\n> `pip install agentic-radar -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple`\n\n### 2. 高级安装（按需）\n根据你使用的智能体框架，安装额外的依赖包以启用完整功能（如预定义工具描述、漏洞测试等）：\n\n*   **CrewAI 用户**：\n    ```sh\n    pip install \"agentic-radar[crewai]\"\n    ```\n\n*   **OpenAI Agents 用户**：\n    ```sh\n    pip install \"agentic-radar[openai-agents]\"\n    ```\n\n## 基本使用\n\nAgentic Radar 主要提供 `scan`（扫描分析）和 `test`（漏洞测试）两个核心命令。\n\n### 1. 扫描工作流并生成报告\n这是最常用的功能，用于分析代码结构、可视化工具调用链并生成 HTML 报告。\n\n**命令格式：**\n```sh\nagentic-radar scan [OPTIONS] FRAMEWORK:{langgraph|crewai|n8n|openai-agents|autogen}\n```\n\n**使用示例：**\n扫描一个 LangGraph 项目文件夹并输出报告：\n```sh\nagentic-radar scan langgraph -i path\u002Fto\u002Flanggraph\u002Fexample\u002Ffolder -o report.html\n```\n*   `-i`: 指定包含智能体代码的目录路径。\n*   `-o`: 指定生成的 HTML 报告文件名。\n\n### 2. (可选) 运行时漏洞测试\n针对 OpenAI Agents 等工作流进行对抗性输入测试（如提示词注入、PII 泄露检测）。\n\n**前置条件：**\n必须设置环境变量：\n```sh\nexport OPENAI_API_KEY=\u003Cyour_api_key>\n```\n\n**使用示例：**\n```sh\nagentic-radar test openai-agents \"examples\u002Fopenai-agents\u002Fbasic\u002Flifecycle_example.py\"\n```\n*   该命令会自动注入测试用例，并在终端以表格形式展示测试结果（通过\u002F失败及原因）。\n\n### 3. (可选) 提示词加固\n自动优化检测到的系统提示词（System Prompts），使其更符合安全最佳实践。\n\n**使用示例：**\n```sh\nagentic-radar scan openai-agents --harden-prompts -i examples\u002Fopenai-agents\u002Fbasic\u002F -o report.html\n```\n*   加固后的提示词将直接显示在生成的 HTML 报告中。\n\n生成报告后，直接在浏览器打开 `report.html` 即可查看详细的工作流图谱、工具列表及安全漏洞映射分析。","某金融科技公司正在开发一套基于多智能体（Multi-Agent）的自动化信贷审批系统，该系统需自主调用外部 API 查询征信、分析流水并生成决策报告。\n\n### 没有 agentic-radar 时\n- **盲区风险高**：开发团队难以直观看到智能体之间复杂的调用链路，无法确认是否存在未授权的工具访问或死循环逻辑。\n- **提示词脆弱**：缺乏针对“提示词注入”攻击的自动化检测，恶意用户可能通过特殊指令诱骗智能体跳过风控规则或直接泄露敏感数据。\n- **人工审计低效**：安全测试依赖人工构造攻击案例，覆盖率低且耗时极长，导致上线前无法全面评估代理工作流的安全隐患。\n- **集成滞后**：安全检查仅在项目末期进行，一旦发现架构级漏洞，往往需要重构大量代码，严重拖慢交付进度。\n\n### 使用 agentic-radar 后\n- **全景可视化**：agentic-radar 自动生成可视化的工作流雷达图，清晰展示每个智能体的权限边界与数据流向，瞬间定位异常调用路径。\n- **主动防御注入**：内置的提示词加固功能自动模拟各类注入攻击，提前识别并修复了可能导致越权操作的指令漏洞。\n- **自动化扫描**：将 agentic-radar 嵌入 CI\u002FCD 流水线，每次代码提交即自动运行深度安全扫描，确保新引入的智能体行为符合安全基线。\n- **左移安全防线**：在开发早期即可发现并解决架构设计缺陷，将原本需要数周的安全加固工作缩短至小时级，保障系统按时上线。\n\nagentic-radar 将原本黑盒且脆弱的智能体工作流转变为透明、可审计且具备内生安全能力的可靠系统。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fsplx-ai_agentic-radar_81c8fb27.png","splx-ai","SplxAI","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fsplx-ai_0013de88.png","GenAI Security Platform",null,"ante@splx.ai","www.splx.ai","https:\u002F\u002Fgithub.com\u002Fsplx-ai",[81,85],{"name":82,"color":83,"percentage":84},"Python","#3572A5",85.2,{"name":86,"color":87,"percentage":88},"Jinja","#a52a22",14.8,950,124,"2026-04-14T01:39:04","Apache-2.0","未说明 (基于 Python，通常支持 Linux, macOS, Windows)","未说明",{"notes":96,"python":97,"dependencies":98},"1. 基础安装仅需 Python (pip)，无其他前置依赖。\n2. 若使用 CrewAI 相关功能（如漏洞测试或工具描述），需安装 'agentic-radar[crewai]'，此时强制要求 Python 版本在 3.10 到 3.13 之间，否则工具描述可能缺失。\n3. 运行漏洞测试功能 ('test' 命令) 需要设置 OPENAI_API_KEY 或 AZURE_OPENAI_API_KEY 环境变量。\n4. 支持扫描的框架包括：LangGraph, CrewAI, n8n, OpenAI Agents, Autogen。",">=3.10 且 \u003C3.13 (安装 CrewAI 扩展时必需); 其他情况仅需安装 Python",[99,100],"crewai-tools (可选，需配合 crewai 扩展)","openai-agents (可选，需配合 openai-agents 扩展)",[35,13,52,15,14],[103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118],"agentic-ai","agentic-framework","agentic-workflow","ai","ai-red-teaming","ai-security","cli","devsecops","llm","llm-security","red-teaming","security","security-tools","generative-ai","mcp","mcp-server","2026-03-27T02:49:30.150509","2026-04-14T20:55:35.449300",[122,127,132,137,142,147],{"id":123,"question_zh":124,"answer_zh":125,"source_url":126},33260,"为什么 Agentic Radar 无法识别 LangGraph 框架中的 Agent，而是将它们全部标记为 BASICS？","这是 LangGraph 框架当前的预期行为。由于 LangGraph 具有动态特性，其代理工作流的定义比其他框架更自由，导致准确检测 Agent 存在困难。开发团队已实施了一种启发式方法（heuristic）来检测 LangGraph 节点是否代表 Agent，这能覆盖许多用例，但受限于框架的复杂性，无法在所有情况下生效。该问题将在未来的版本中持续改进。","https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F89",{"id":128,"question_zh":129,"answer_zh":130,"source_url":131},33261,"在分析 n8n 工作流时遇到“not well-formed (invalid token)”错误或节点标签中包含特殊字符导致解析失败，如何解决？","此问题通常由 HTML 中未转义的字符引起。请升级到最新版本（v0.4.1 或更高），该版本已修复了有关 HTML 中未转义字符的问题。此外，请注意该工具旨在针对包含单个工作流的目录生成一份综合报告，因此如果目录中包含多个工作流，生成的图表和报告可能会显得拥挤，且不同工作流的节点可能会混合显示漏洞信息。","https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F47",{"id":133,"question_zh":134,"answer_zh":135,"source_url":136},33262,"在 Windows 上生成的报告中，图表节点图标缺失并提示\"No loadimage plugin for svg:cairo\"警告，该如何解决？","这是因为 Windows 环境下 Graphviz 缺少必要的图像加载插件支持。确保您已从官网下载并安装了 Graphviz，同时需要安装 Cairo（通常作为 GTK 的一部分）。如果问题仍然存在，请检查是否已应用相关修复补丁（如 PR #51），该补丁旨在解决 Windows 上的图标渲染问题。","https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F44",{"id":138,"question_zh":139,"answer_zh":140,"source_url":141},33263,"在 Windows 上生成的报告中，列表项前的圆点符号\"•\"显示为乱码\"â€\"，如何修复？","这是由于 Windows 上的字符编码问题导致的。当生成报告时，UTF-8 编码的特殊字符未被正确解析。建议检查生成环境的编码设置，确保终端和文件系统均使用 UTF-8 编码。此问题已在后续更新中被关注并修复，请尝试更新到最新版本以获取正确的字符渲染。","https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F45",{"id":143,"question_zh":144,"answer_zh":145,"source_url":146},33264,"如何使用自定义配置文件运行 `agentic-radar test` 命令？","您可以使用 `--config` 参数导入自定义测试配置。配置文件应为 YAML 格式，包含一个 `tests` 列表，每个测试项需定义 `name`（名称）、`input`（输入）和 `success_condition`（成功条件）。例如：\n```yaml\ntests:\n   - name: TestA\n     input: Mention A.\n     success_condition: Attack succeeded if A is mentioned.\n```\n此外，您可以在配置中添加 `include_default_tests: true` 以同时运行默认测试套件，若设为 `false` 则仅运行用户提供的测试。","https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F82",{"id":148,"question_zh":149,"answer_zh":150,"source_url":151},33265,"Agentic Radar 能否提取 OpenAI Agents 或 CrewAI 代理的元数据（如使用的模型和系统提示词）？","是的，该功能已被纳入开发计划。对于 OpenAI Agents，工具将解析 Agent 构造函数中的 `model` 参数以获取使用的 LLM 模型，并从 `instructions` 参数中提取系统提示词。对于 CrewAI，工具将从 Agent 构造函数、YAML 配置文件以及源代码中解析模型信息，并提取用户传递的提示模板作为系统提示词。这些元数据将被添加到输出报告中，以提高工作流的透明度并帮助检测潜在的系统提示词漏洞。","https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F65",[153,158,163,168,173,178,183,188,193,198,203,208,213,218,222,226,230,234,239,243],{"id":154,"version":155,"summary_zh":156,"released_at":157},255425,"v0.14.1","## [0.14.1](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.14.0...v0.14.1) (2025-11-27)\n\n\n### 错误修复\n\n* **openai-agents:** 防止图定义中出现重复的工具节点 ([#118](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F118)) ([eb3184d](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002Feb3184d5a3fa7f65b9a7e03035ca716a501d04a9))\n* 更新存在漏洞的依赖项 ([#120](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F120)) ([fc6788e](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002Ffc6788ef18711510a88bb5f9b9e783b8b07250a1))","2025-11-27T15:28:44",{"id":159,"version":160,"summary_zh":161,"released_at":162},255426,"v0.14.0","## [0.14.0](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.13.0...v0.14.0) (2025-10-08)\n\n\n### 功能特性\n\n* **autogen:** 为 autogen agentchat 添加 MCP 检测 ([#111](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F111)) ([902a4cd](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F902a4cd2736d05b799df8719fd86c475b7d318df))\n* **crewai:** 增加对 crewai MCP 服务器检测的支持 ([#112](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F112)) ([9e1f302](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F9e1f302fb510b811aada992b8855c0cd3b24c484))\n* **n8n:** 在 n8n 中添加 MCP 检测 ([#114](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F114)) ([f3e7035](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002Ff3e7035b7abbed53fdd7ac8f6ce0e74237fce579))\n\n\n### Bug 修复\n\n* **openai-agents:** 提升 MCP 检测的覆盖范围 ([#113](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F113)) ([036ed80](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F036ed80fc963434c6785c5d096e8de8e9b428d29))","2025-10-09T08:10:23",{"id":164,"version":165,"summary_zh":166,"released_at":167},255427,"v0.13.0","## [0.13.0](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.12.0...v0.13.0) (2025-07-17)\n\n\n### 功能\n\n* 添加 GitHub 工作流示例 YAML ([#101](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F101)) ([e3be24f](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002Fe3be24fa9f9df0ce55d57e21815d521347a6de5b))\n\n\n### 错误修复\n\n* **autogen:** 在报告中显示无团队的智能体 ([#103](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F103)) ([22d79bf](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F22d79bf1fd26e1d27e1b9b6e6c60f5568f3e6658))","2025-07-17T14:36:31",{"id":169,"version":170,"summary_zh":171,"released_at":172},255428,"v0.12.0","## [0.12.0](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.11.1...v0.12.0) (2025-06-09)\n\n\n### 功能\n\n* 将图导出为 JSON 格式 ([#99](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F99)) ([30f6c89](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F30f6c89ad86340c4d825cd9f193046ebc82badae))","2025-06-17T12:31:59",{"id":174,"version":175,"summary_zh":176,"released_at":177},255429,"v0.11.1","## [0.11.1](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.11.0...v0.11.1) (2025-06-03)\n\n\n### 错误修复\n\n* 将报告表格中的字体颜色设置为黑色 ([#96](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F96)) ([a81c92b](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002Fa81c92b3188fc5644a589eb7d722e0bc123cd8a6))","2025-06-03T15:22:32",{"id":179,"version":180,"summary_zh":181,"released_at":182},255430,"v0.11.0","## [0.11.0](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.10.1...v0.11.0) (2025-06-03)\n\n\n### 功能\n\n* 支持 Autogen AgentChat 扫描 ([#95](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F95)) ([13edf17](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F13edf17ff90c24e8887e014b4419f25a296f5d0f))\n* **langgraph:** 实现用于检测智能体的简单启发式方法 ([#92](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F92)) ([aa4e86a](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002Faa4e86aaa9294feb4d8a7a1f482473a785655ed0))","2025-06-03T12:43:51",{"id":184,"version":185,"summary_zh":186,"released_at":187},255431,"v0.10.1","## [0.10.1](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.10.0...v0.10.1)（2025-05-23）\n\n\n### 错误修复\n\n* 缺少与 pyyaml 和 openai-agents 相关的依赖项（[#90](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F90)）（[16e3f5b](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F16e3f5b35b743569a02143ab8339094255bcd884)）","2025-05-23T15:41:16",{"id":189,"version":190,"summary_zh":191,"released_at":192},255432,"v0.10.0","## [0.10.0](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.9.1...v0.10.0) (2025-05-12)\n\n\n### 功能特性\n\n* 为 LangGraph 添加 MCP 支持的初始提交 ([#78](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F78)) ([5573d83](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F5573d834fcaa5f7c3b68f5ba58c8b0e36540eead))\n* **提示加固：** 添加 PII 保护步骤 ([#85](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F85)) ([574c859](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F574c8596bc86830dc546f7d69003b3c7395f2821))\n* **雷达测试：** 为测试添加可选配置 ([#83](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F83)) ([2d3c61e](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F2d3c61ebdcc94ad9870e00b8145ff00ad1fcd2bc))\n* 将“探针”重命名为“测试” ([#80](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F80)) ([153a418](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F153a4180c8073f0020c73174408bdd1bec767bdb))\n* 将“提示增强”重命名为“提示加固” ([#84](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F84)) ([fcb941f](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002Ffcb941f44a31e9eed008f060d067cd0ecd3778f9))","2025-05-13T07:37:29",{"id":194,"version":195,"summary_zh":196,"released_at":197},255433,"v0.9.1","## [0.9.1](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.9.0...v0.9.1) (2025-04-26)\n\n\n### 功能\n\n* **升级：** 临时添加 openai-agents 依赖 ([#76](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F76)) ([fc5cac2](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002Ffc5cac2a2cdc32ccb6fa5d6f53941a946efb17f7))","2025-04-26T07:40:11",{"id":199,"version":200,"summary_zh":201,"released_at":202},255434,"v0.9.0","## [0.9.0](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.8.0...v0.9.0) (2025-04-26)\n\n\n### 功能\n\n* **升级：** 发布 0.8.1 ([#74](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F74)) ([e6c538f](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002Fe6c538fea22067961e870c17f7bb0858362c870f))","2025-04-26T07:28:35",{"id":204,"version":205,"summary_zh":206,"released_at":207},255435,"v0.8.0","## [0.8.0](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.7.0...v0.8.0) (2025-04-26)\n\n\n### Features\n\n* **agent-vulnerability:** OpenAI Agents Improvements ([95bb67f](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F95bb67f0b4c7844c2d82985aa2ad67cca59d3c82))\n* **probe:** Init Agentic Probe ([2e72045](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F2e72045703edae9cd97cf3b092a4599680901ad5))","2025-04-26T06:39:53",{"id":209,"version":210,"summary_zh":211,"released_at":212},255436,"v0.7.0","## [0.7.0](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.6.0...v0.7.0) (2025-04-23)\n\n\n### Features\n\n* adds initial unit tests for the LangGraph framework ([23dd317](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F23dd317fc0798dc5378f9fd99b23cd1d0438e4e8))\n* **crewai:** add additional agent information to report ([#64](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F64)) ([2e4e2b2](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F2e4e2b29fa55c9bcae7a29c557d0c595aa27b554))\n* **openai-agents:** add additional agent information to report ([#66](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F66)) ([0c880ab](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F0c880ab4179d8653504bde4f391c236da8f738c6))\n* **openai-agents:** detection of MCP servers ([#68](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F68)) ([42a64b6](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F42a64b6fb8ce36adf6f1cb151cb25725670d92b0))\n* Prompt Enhancement ([#67](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fissues\u002F67)) ([39b45e4](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F39b45e4d13e1c272e4c59f176580fa4ab4e7f358))","2025-04-23T15:32:06",{"id":214,"version":215,"summary_zh":216,"released_at":217},255437,"v0.6.0","## [0.6.0](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcompare\u002Fv0.5.1...v0.6.0) (2025-03-28)\n\n\n### Features\n\n* add OpenAI Agents framework analyzer ([6d20adb](https:\u002F\u002Fgithub.com\u002Fsplx-ai\u002Fagentic-radar\u002Fcommit\u002F6d20adb95a34d4738a462ead5002dafedf2a5281))","2025-03-29T18:32:45",{"id":219,"version":220,"summary_zh":76,"released_at":221},255438,"v0.5.1","2025-03-25T14:50:17",{"id":223,"version":224,"summary_zh":76,"released_at":225},255439,"v0.5.0","2025-03-25T11:43:13",{"id":227,"version":228,"summary_zh":76,"released_at":229},255440,"v0.4.1","2025-03-24T13:39:24",{"id":231,"version":232,"summary_zh":76,"released_at":233},255441,"v0.4.0","2025-03-24T11:32:19",{"id":235,"version":236,"summary_zh":237,"released_at":238},255442,"v0.3.2","## What's Changed\r\n* fix: add encoding for the vulnerabilities JSON (Fixes #45)\r\n","2025-03-21T14:28:29",{"id":240,"version":241,"summary_zh":76,"released_at":242},255443,"v0.3.1","2025-03-20T14:17:23",{"id":244,"version":245,"summary_zh":76,"released_at":246},255444,"v0.3.0","2025-03-18T07:19:52"]