Agent Sandbox
技能 已验证 活跃Agent skill for sandbox - invoke with $agent-sandbox
To provide developers and testers with a secure and efficient way to create, manage, and utilize isolated execution environments for code development and testing.
功能
- Create and configure E2B sandboxes
- Execute code safely in isolated environments
- Manage sandbox lifecycles (creation to termination)
- Handle file uploads, downloads, and environment configuration
- Monitor sandbox performance and resource utilization
使用场景
- Setting up isolated development environments for new projects
- Running code tests in a clean, reproducible environment
- Experimenting with new libraries or frameworks without affecting the local system
- Managing and cleaning up temporary execution environments after use
非目标
- General-purpose code execution outside of a managed sandbox
- Managing cloud infrastructure beyond E2B sandbox instances
- Performing security audits of user-submitted code within the sandbox
Documentation
- info:Configuration & parameter referenceThe SKILL.md provides examples of tool usage with parameters, but a formal reference for all options and defaults is not explicitly laid out.
Code Execution
- info:ValidationWhile the code examples show structured parameters, there's no explicit mention of using a schema validation library like Zod or Pydantic for input sanitization.
- info:Error HandlingThe SKILL.md outlines core responsibilities including 'Proper error handling and logging', but specific details on structured error reporting or fail-closed behavior are not provided.
- info:LoggingThe SKILL.md mentions 'Clear execution logging and debugging support' as a quality standard, but details on a specific audit file or structured logging format are absent.
Errors
- info:Actionable error messagesWhile the SKILL.md mentions proper error handling, specific details on how errors are presented to the user with remediation steps are not fully detailed.
Protocol
- info:Idempotent retry & timeoutsThe `timeout` parameter is shown for `sandbox_create`, suggesting some consideration for timeouts, but idempotency and structured retryable errors are not explicitly detailed.
Practical Utility
- info:Edge casesThe SKILL.md mentions quality standards like 'Implement proper error handling and logging' and 'Scalable sandbox orchestration', but explicit documentation of failure modes and recovery steps for edge cases is not detailed.
Safety
- info:Halt on unexpected stateThe skill aims to manage sandboxes, and while error handling is mentioned, explicit documentation on halting on unexpected pre-state with clear reports is not present.
安装
npx skills add ruvnet/ruflo通过 npx 运行 Vercel skills CLI(skills.sh)— 需要本地安装 Node.js,以及至少一个兼容 skills 的智能体(Claude Code、Cursor、Codex 等)。前提是仓库遵循 agentskills.io 格式。
质量评分
已验证类似扩展
Minimal Run And Audit
100用于 README 优先的 AI 代码库进行可信执行和报告的技能。当任务是专门从选定的 smoke test 或已文档化的推理或评估命令捕获或标准化证据，并写入标准化的 `repro_outputs/` 文件（包括在存储库文件更改时生成补丁说明）时使用。请勿用于训练执行、初始代码库引入、通用环境设置、论文查找、目标选择或单独的端到端编排。
Wasm Agent
99Create and manage sandboxed WASM agents for isolated code execution
Python Executor
99Execute Python code in a safe sandboxed environment via [inference.sh](https://inference.sh). Pre-installed: NumPy, Pandas, Matplotlib, requests, BeautifulSoup, Selenium, Playwright, MoviePy, Pillow, OpenCV, trimesh, and 100+ more libraries. Use for: data processing, web scraping, image manipulation, video creation, 3D model processing, PDF generation, API calls, automation scripts. Triggers: python, execute code, run script, web scraping, data analysis, image processing, video editing, 3D models, automation, pandas, matplotlib
Cloudflare Sandbox SDK
99构建用于安全代码执行的沙盒应用程序。在构建 AI 代码执行、代码解释器、CI/CD 系统、交互式开发环境或执行不受信任的代码时加载。涵盖 Sandbox SDK 生命周期、命令、文件、代码解释器和预览 URL。倾向于从 Cloudflare 文档中检索信息，而不是依赖预训练知识。
AgentBay AIO 技能
95用于任何“运行/执行/评估代码”的请求（如 run this code、execute script、用 Python 画图、跑代码/画图/执行脚本）。通过 AgentBay SDK 创建 code_latest 沙箱，调用 run_code 执行并返回结果。支持 Python、JavaScript、R、Java。
Context Mode
100从 GitHub 更新 context-mode 并修复 hooks/settings。 拉取最新代码，构建，安装，更新 npm 全局包，配置 hooks。 触发器：/context-mode:ctx-upgrade