[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-79941":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":19,"compositeScore":20,"rankGlobal":10,"rankLanguage":10,"license":10,"archived":21,"fork":21,"defaultBranch":22,"hasWiki":21,"hasPages":21,"topics":23,"createdAt":10,"pushedAt":10,"updatedAt":29,"readmeContent":30,"aiSummary":31,"trendingCount":16,"starSnapshotCount":16,"syncStatus":32,"lastSyncTime":33,"discoverSource":34},79941,"Jumping-Agent-platform","answeryt\u002FJumping-Agent-platform","answeryt","Build your own AI agent through gameplay.","",null,"TypeScript",215,51,41,1,0,43,76,129,92.15,false,"main",[24,25,26,27,28],"agent","agentic-workflow","ai-agent","platform","workflow","2026-06-12 04:01:25","\u003Cdiv align=\"center\">\n  \u003Cimg src=\"8315c2b3b7fabe9b0e95a8c6c7b4f9db.jpg\" alt=\"Jumping Agent: a super simple Agent building platform (Task split -> Community \u002F Code \u002F Data agent -> Aggregator)\" width=\"90%\">\n  \u003Ch1>Jumping-Agent\u003C\u002Fh1>\n  \u003Cp>\u003Cstrong>Turn flat Agent building into spatial Agent building through gameplay—easy for beginners to get started\u003C\u002Fstrong>\u003C\u002Fp>\n  \u003Cp>\n    \u003Ca href=\"README_EN.md\">English\u003C\u002Fa> |\n    \u003Ca href=\"README.md\">中文\u003C\u002Fa>\n  \u003C\u002Fp>\n\u003C\u002Fdiv>\n\n## Project Overview\n\nJumping Agent is a fun and approachable project that uses a \"jumping game\" style interaction to help beginners build their own Agents on mobile devices, especially tablets. It turns abstract Agent workflows into visible, touchable, step-by-step jumps, making the whole process easier to understand and use.\n\nInstead of showing users a flat workflow full of complex lines and arrows, this project presents the workflow in a spatial and game-like way. By pressing the screen and controlling a character to jump forward, users can intuitively understand how an Agent workflow moves from one step to the next.\n\n![Jumping-style Agent builder interface: press to jump and understand how the workflow moves forward](b27fe65904b078f84f58a00f6d6e7488.jpg)\n\n## Architecture\n\n- **`Frontend\u002F`**: The jumping-style Agent building interface. It handles workflow display, node configuration, and the chat entry point.\n- **`agent_builder\u002F`**: Agent skeletons and templates, including workflow templates, Agent templates, project templates, and configuration generation logic.\n- **`back_agent\u002F`**: Reads skeleton code and uses the user's frontend input to complete, modify, and generate the Agent implementation.\n- **`backend\u002F`**: The orchestration service. `orchestrator.py` connects frontend requests, dynamically loads `agent_builder`, generates Agent workspaces, and calls `back_agent` through local HTTP requests.\n- **`sandbox\u002F` \u002F `sandbox-main\u002F`**: Provides sandbox tool capabilities for generated Agents, such as file operations, shell commands, and browser automation. This helps prevent Agents from operating directly on the host machine. Many thanks to the provider of this sandbox project: [agent-infra\u002Fsandbox](https:\u002F\u002Fgithub.com\u002Fagent-infra\u002Fsandbox).\n\nDefault service URLs:\n\n- Frontend: `http:\u002F\u002Flocalhost:6301`\n- back_agent: `http:\u002F\u002Flocalhost:8000\u002Fchat`\n- backend \u002F Orchestrator: `http:\u002F\u002Flocalhost:8001`\n- Sandbox: `http:\u002F\u002Flocalhost:8080`\n\n## Quick Start\n\n### Prerequisites\n\n- Git\n- Python 3.11+\n- Node.js 18+\n- npm\n- Docker\n\n### Clone\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fansweryt\u002FFat-Cat-Agent-platform-1.0.git\ncd Jumping-Agent-platform-1.0\n```\n\n### Install Frontend Dependencies\n\n```bash\ncd Frontend\nnpm install\ncd ..\n```\n\n### Install Backend Dependencies\n\nIt is recommended to create a virtual environment first:\n\n```bash\npython -m venv .venv\n```\n\nWindows:\n\n```bash\n.venv\\Scripts\\activate\n```\n\nmacOS \u002F Linux:\n\n```bash\nsource .venv\u002Fbin\u002Factivate\n```\n\nInstall Python dependencies:\n\n```bash\npython -m pip install \"fastapi\" \"uvicorn[standard]\" \"pydantic\" \"openai\"\n```\n\n### Configure API Key\n\n`back_agent\u002Fconfig\u002Fmodel_config.toml` reads the `OPENAI_API_KEY` environment variable by default. You can set the environment variable directly, or use the built-in project script to write it into `.env`.\n\nOption 1: use an environment variable.\n\nWindows PowerShell:\n\n```powershell\n$env:OPENAI_API_KEY=\"your_api_key_here\"\n```\n\nmacOS \u002F Linux:\n\n```bash\nexport OPENAI_API_KEY=\"your_api_key_here\"\n```\n\nOption 2: use the project script.\n\n```bash\npython backend\u002Fset_agent_api_key.py\n```\n\nThe script will ask for your API key and update `back_agent\u002F.env` as well as the `.env` files in generated workspaces.\n\n### Start Sandbox\n\n```bash\ndocker run --security-opt seccomp=unconfined --rm -it -p 8080:8080 ghcr.io\u002Fagent-infra\u002Fsandbox:latest\n```\n\n### Start Services\n\nOpen 3 terminals and run each command from the repository root.\n\nTerminal A: start `back_agent`.\n\n```bash\ncd back_agent\npython -m uvicorn api:app --host 0.0.0.0 --port 8000\n```\n\nTerminal B: start `backend` \u002F Orchestrator.\n\n```bash\ncd backend\npython -m uvicorn orchestrator:app --host 0.0.0.0 --port 8001\n```\n\nTerminal C: start the frontend.\n\n```bash\ncd Frontend\nnpm run server -- --host 0.0.0.0 --port 6301 --allowed-hosts all\n```\n\nOpen the browser and visit:\n\n```text\nhttp:\u002F\u002Flocalhost:6301\n```\n\n## iPad \u002F LAN Access\n\nIf you want to access the frontend from an iPad, make sure the iPad and your computer are on the same local network, then visit the computer's LAN IP address.\n\nCheck the IP address on Windows:\n\n```powershell\nipconfig\n```\n\nCheck the IP address on macOS \u002F Linux:\n\n```bash\nifconfig\n```\n\nFind the IPv4 address of your current network adapter, for example `192.168.x.x`, then open this address in the iPad browser:\n\n```text\nhttp:\u002F\u002F192.168.x.x:6301\n```\n\n## CLI Plan\n\nThe repository can currently be run with the manual source commands above. To make it easier for GitHub users to get started, a CLI could later wrap these steps, for example:\n\n```bash\nagent-jump setup\nagent-jump config set OPENAI_API_KEY\nagent-jump dev\n```\n\nIdeally:\n\n- `agent-jump setup` installs frontend and backend dependencies.\n- `agent-jump config set OPENAI_API_KEY` writes the model API key.\n- `agent-jump dev` starts Sandbox, back_agent, backend, and Frontend together.\n\nBefore the CLI is officially implemented, please use the manual startup steps in `Quick Start`.\n\n## Acknowledgements\n\nThis project was independently developed by me. The sandbox capability is based on [agent-infra\u002Fsandbox](https:\u002F\u002Fgithub.com\u002Fagent-infra\u002Fsandbox), and I sincerely appreciate their work.\n\nDue to limited personal time and experience, the current version still has some limitations. For example, sandbox operations may occasionally fail, the Agent workflow currently provides only 7 templates, and the jump-platform orchestration plus final build process are not yet fully stable. I will continue improving the project.\n\nIf you would like to help make this project better, issues and pull requests are welcome. You can also contact me directly at [answeryt@qq.com](mailto:answeryt@qq.com). Chinese users may contact me on WeChat: `answerYTAarun`.\n\nI also hope more beginners can use it to explore their imagination and build their own Agents.\n","Jumping-Agent 是一个通过游戏化方式帮助初学者构建AI代理的平台。其核心功能是将复杂的代理工作流转化为直观易懂的“跳跃”式交互，用户可以通过触摸屏幕控制角色前进，从而轻松理解并创建自己的代理流程。技术上，该项目使用Python开发，并结合了前端界面、后端服务及沙盒工具来实现整个代理构建过程的安全与便捷。适用于希望以更有趣且直观的方式学习和实践AI代理构建的个人或团队，特别适合在平板等移动设备上操作。",2,"2026-06-11 03:58:36","CREATED_QUERY"]