skills.rest
ruvnet
ruvnetCOMMUNITY
Explore all skills by this author →
Website𝕏@ruvGitHub

wasm-agent

Create and manage sandboxed WebAssembly agents for isolated code execution.

70.1k|8.4k|Updated Jun 2, 2025
Search Tags:#mcp#code execution#webassembly#Wasm#Agent#Sandbox#Isolation
One-click install
npx skills add https://github.com/ruvnet/claude-flow --skill wasm-agent
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: wasm-agent
Source: https://github.com/ruvnet/claude-flow/tree/main/plugins/ruflo-agent/skills/wasm-agent
Command: npx skills add https://github.com/ruvnet/claude-flow --skill wasm-agent

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Running untrusted or experimental agent code directly on a host machine risks filesystem damage and unpredictable behavior. This Skill lets you spin up fully sandboxed agents inside WebAssembly runtimes so code executes in isolation with no access to the host system.

Core Features & Use Cases

  • Sandboxed Agent Creation: Create WASM-based agents that run in complete isolation from the host filesystem.
  • Interactive Agent Management: Send prompts, grant tool access, and read or write files inside the agent sandbox.
  • Portable Export: Package agents for sharing and run them on any WASM-compatible runtime with reproducible behavior.
  • Use Case: You want to test a new agent configuration without risking your development environment. Create a WASM agent, send it prompts, inspect its sandboxed files, and terminate it when finished.

Quick Start

Ask the assistant to create a new WASM agent with a given configuration and then send it a prompt to verify it responds inside the sandbox.

Frequently Asked Questions about wasm-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a sandboxed WASM agent?▼

Call the wasm_agent_create MCP tool with your agent configuration to instantiate a sandboxed agent. Once created, you can send prompts, attach tools, and manage its files entirely within the isolated environment.

How do I run untrusted code safely with WebAssembly?▼

WebAssembly sandboxes isolate execution from the host filesystem, so untrusted code cannot damage your system. Create a WASM agent, execute the code inside it, and terminate the agent when finished.

Can I export a WASM agent to run on other platforms?▼

Yes, the wasm_agent_export tool packages an agent for sharing. The exported agent runs on any WASM-compatible runtime with reproducible behavior across platforms.

Can a WASM agent access my host filesystem?▼

No, WASM agents run in full sandbox isolation and cannot access the host filesystem. File operations are confined to the sandbox through the wasm_agent_files tool.

How do I stop or remove a running WASM agent?▼

Use the wasm_agent_terminate tool to stop a specific agent. You can first call wasm_agent_list to see all currently running WASM agents and identify the one to terminate.

Related Skills

Explore More Software Engineering Skills→
Zhow01Zhow01
52

moltguard

Detect and block unsafe prompts and secret leakage in OpenClaw deployments.

Community
Intermediate
Uomi-networkUomi-network

wasp-deploy

Deploy compiled UOMI WASM agents to the UOMI network and mint them as NFTs.

Official
Advanced
Motium-AIMotium-AI

harness-test

Run Claude Code harness tests in an isolated sandbox.

Official
Advanced
OpenHandsOpenHands
134

docker

Start a Docker daemon and run containers inside containerized environments.

Official
Advanced