← 返回 Skills 市场

E2b Sandbox Runtime

Name: E2b Sandbox Runtime
Author: tangweigang-jpg

作者 Tang Weigang · GitHub ↗ · v0.1.0 · MIT-0

cross-platform ⚠ suspicious

总下载

当前安装

版本数

在 OpenClaw 中安装

/install e2b-sandbox-runtime

功能描述

E2B：在隔离 micro-VM 里执行 AI 生成代码的云端 runtime。Python / TS SDK 通过 Connect-RPC 调用 envd 守护进程（Rust + protobuf）。 E2B: cloud-side runtime for executing AI-generated code...

使用说明 (SKILL.md)

这个 skill 适合什么用户？能做哪些任务？

概览

E2B 是在隔离 micro-VM 里执行 AI 生成代码的云端 runtime（github.com/e2b-dev/E2B）。Python / TypeScript SDK 是纯 RPC 客户端，通过 Connect-RPC 对接 envd 守护进程（Rust + protobuf，托管在独立的 e2b-dev/infra repo）。

SDK 接口为 2x2：{Sandbox, AsyncSandbox} × {Template, AsyncTemplate}。Sandbox / AsyncSandbox 各带四个子模块作为实例属性：files: Filesystem / comman...

Doramagic 晶体页: https://doramagic.ai/zh/crystal/e2b-sandbox-runtime

知识规模

42 条约束 (3 fatal + 39 non-fatal)
上游源码: e2b-dev/E2B @ commit 557b723c
蓝图 ID: finance-bp-141

用法

Host AI（Claude Code / Cursor / OpenClaw）读 references/seed.yaml，按其中的：

intent_router 匹配用户意图
architecture 理解项目架构
constraints 应用 anti-pattern 约束
business_decisions 参考核心设计决策

FAQ 摘要

这个 skill 适合什么用户？能做哪些任务？

适合需要让 LLM 安全执行任意代码的工程师：AI 助手跑 Python / shell、数据分析 sandbox、教学环境。每个 sandbox 是隔离 micro-VM，泄露面限制在单个 sandbox 内。Jupyter 内核 / 富输出场景应改用 e2b-code-interpreter。访问 doramagic.ai/r/e2b 查看完整用例。

需要准备什么环境？依赖什么？

Python 3.9+ 或 Node 18+。E2B_API_KEY 给 SDK runtime（创建 / 控制 sandbox）；E2B_ACCESS_TOKEN 给 CLI 鉴权（template build / dashboard）。可选 E2B_DOMAIN（默认 e2b.app）。出站 HTTPS 到 *.e2b.app。

会踩哪些坑？这个 skill 怎么防护？

本 skill 内置 42 条约束（3 条 fatal）。典型踩坑：(1) TS 用户忘 kill()——TS Sandbox 没有 Symbol.dispose 也没自动清理，sandbox 持续计费到超时；(2) 通过 envs={...} 传 API key 会被 commands.list() 泄露给持有 sandbox 句柄的人；

完整文档: 见 references/seed.yaml (v6.1 schema). 浏览页: https://doramagic.ai/zh/crystal/e2b-sandbox-runtime

安全使用建议

This skill appears to be a knowledge/blueprint for an E2B sandbox runtime but its metadata omits credentials and filesystem/egress expectations that the SKILL.md/seed.yaml explicitly describe. Before installing or enabling it: (1) ask the publisher for the homepage/repository and for a clear list of required environment variables (E2B_API_KEY, E2B_ACCESS_TOKEN, E2B_DOMAIN) to be declared in registry metadata; (2) do not provide real production API keys until you confirm the author and trust the runtime — use a limited or test key instead; (3) be aware the instructions tell the agent to run local commands, inspect ZVT_HOME and other paths, and may start long-running servers/ports — run first in an isolated/test environment; (4) ask the publisher to update the skill metadata to declare required env vars and any workspace write operations, and to justify why the agent must re-read seed.yaml and run the listed preconditions; (5) if you are uncomfortable with autonomous actions, disable autonomous invocation or restrict the skill until provenance and required privileges are clarified.

功能分析

Type: OpenClaw Skill Name: e2b-sandbox-runtime Version: 0.1.0 The skill bundle provides a runtime for the E2B sandbox and ZVT quant trading library. It is exceptionally security-conscious, containing numerous explicit constraints and warnings (e.g., e2b-C-002, e2b-C-029, and e2b-C-033 in references/seed.yaml) designed to prevent credential leakage, shell injection, and unauthorized network egress. The high-risk capabilities, such as shell execution and network access, are inherent to the stated purpose of an isolated code execution environment and are accompanied by defensive instructions for the AI agent.

能力标签

cryptocan-make-purchasesrequires-oauth-tokenrequires-sensitive-credentials

能力评估

⚠ Purpose & Capability

The skill is presented as a cloud-side sandbox runtime knowledge artifact, which can plausibly require control-plane credentials (E2B_API_KEY) and domain access. However the registry metadata lists no required environment variables or config paths while SKILL.md and seed.yaml explicitly reference E2B_API_KEY, E2B_ACCESS_TOKEN, optional E2B_DOMAIN, and outbound HTTPS to *.e2b.app. That discrepancy (declaring none but documenting secrets) is incoherent and unexplained.

⚠ Instruction Scope

SKILL.md and the included seed.yaml instruct the host AI to re-read seed.yaml on every behavioral decision, run precondition shell/python checks (e.g., import zvt, check ZVT_HOME and file-system write permissions), and follow an execution protocol that writes/uses host_workspace scripts/skills/trace paths. The instructions therefore direct the agent to run local commands, inspect environment and filesystem, and (per human_summary) handle long-running servers and port exposure — actions broader than a passive 'knowledge' skill and not declared in metadata.

✓ Install Mechanism

No install spec or code files are present; this is instruction-only. That lowers installation risk because nothing is downloaded or written by the skill package itself.

⚠ Credentials

Metadata shows no required env vars and sets primaryEnv to the nonstandard 'knowledge', yet SKILL.md documents E2B_API_KEY and E2B_ACCESS_TOKEN (and optional E2B_DOMAIN) and warns about passing API keys into sandboxes. Requiring API keys and telling the agent to run local checks and potentially expose ports is a materially higher privilege request than what the registry declares.

ℹ Persistence & Privilege

The skill does not request 'always: true' and leaves autonomous invocation enabled (platform default). It expects the host to use workspace paths (scripts_path, skills_path, .trace/), which implies the agent will create/read workspace files during execution. That behavior is plausible for this blueprint but should have been declared up-front; combined with missing credential declarations this increases risk.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install e2b-sandbox-runtime
安装完成后，直接呼叫该 Skill 的名称或使用 /e2b-sandbox-runtime 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v0.1.0

E2B AI code sandbox skill — 42 constraints / 3 fatal. 2x2 surface (Sandbox/AsyncSandbox × Template/AsyncTemplate).

元数据

Slug e2b-sandbox-runtime

版本 0.1.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题