Build sandboxed code execution environments on Cloudflare Workers using the Sandbox SDK.
npx skills add https://github.com/vedang/dotagents --skill sandbox-sdk-vedang
Please help me install this Agent Skill. Skill: sandbox-sdk Source: https://github.com/vedang/dotagents/tree/main/specific_skills/sandbox-sdk Command: npx skills add https://github.com/vedang/dotagents --skill sandbox-sdk-vedang
exec() or execute Python, JavaScript, and TypeScript with runCode() for rich outputs and persistent state. - File and Port Management: Write, read, and list files inside sandboxes, and expose HTTP services via preview URLs. - Use Case: Build an AI code interpreter where an agent generates Python data analysis code, executes it in an isolated per-session sandbox with pandas, and returns charts and results to the user. ## Quick Start Ask the agent to create a Cloudflare Worker that uses the Sandbox SDK to run Python code in an isolated sandbox and return the result.High-intent search queries and answers about installing and using this skill.
Use sandbox.runCode() with a code context created via sandbox.createCodeContext(). It supports Python, JavaScript, and TypeScript, returns rich outputs like text and images, and persists state within the context across executions.
Call sandbox.exec() with the command string, optionally passing cwd, env, timeout, and stdin. It returns stdout, stderr, exitCode, and a success flag, making it suitable for build and test pipelines.
You need a containers entry with class_name Sandbox and a Dockerfile image, a Durable Object binding named Sandbox, and a migration with new_sqlite_classes. The Worker entry must also re-export the Sandbox class.
Yes, sandbox.exposePort() returns a preview URL for HTTP services running in the sandbox. In production this requires a custom domain with wildcard DNS, since .workers.dev does not support preview URL subdomains.
A common cause is missing the required export of the Sandbox class from the Worker entry file. Also verify the wrangler containers, Durable Object bindings, and migrations match the required structure exactly.
Use exec() for shell commands, scripts, and build pipelines where you need exit codes and stderr. Use runCode() for LLM-generated code and data analysis where you want rich outputs and persistent state across calls.
cloudflareDeploy and manage Cloudflare Workers and related resources via the wrangler CLI.
dwhobanExecute Python, JavaScript, and TypeScript code in isolated Cloudflare Workers sandboxes.
rnavarychImplement JWT, OAuth/OIDC, session, and RBAC/ABAC authentication patterns.
superserve-aiRun AI code in Firecracker microVM sandboxes with persistent environments.