Wasm Agent
- 645 installs
- 67k repo stars
- Updated August 4, 2026
- ruvnet/ruflo
wasm-agent is a Claude Code skill that creates and controls sandboxed WebAssembly agents for developers who need isolated, portable code execution inside Claude, Cursor, or Codex sessions.
About
wasm-agent is a Claude Code skill that provisions sandboxed WebAssembly agents for safe, isolated execution without leaving the IDE. Four primary commands—create, list, prompt, and terminate—drive seven claude-flow wasm_agent MCP tools covering agent creation, prompting, tool invocation, file access, export, and teardown. Developers reach for wasm-agent when running untrusted code, experimenting with agent configurations, or building portable agents that run anywhere WASM is supported. The workflow keeps experimentation contained while still allowing prompt-driven iteration and artifact export from the sandbox.
- Execute WASM binaries as native agent tools inside LLM sessions
- Zero-install lightweight compute for agentic workflows
- Runs in sandbox with filesystem and network permissions
- Compatible with Rust, Go, Python, and AssemblyScript compiled modules
- Reduces token usage and latency versus full container calls
Wasm Agent by the numbers
- 645 all-time installs (skills.sh)
- +10 installs in the week ending Jul 26, 2026 (Skillselion tracking)
- Ranked #1,500 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
- Data as of Aug 5, 2026 (Skillselion catalog sync)
npx skills add https://github.com/ruvnet/ruflo --skill wasm-agentAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 645 |
|---|---|
| repo stars | ★ 67k |
| Last updated | August 4, 2026 |
| Repository | ruvnet/ruflo ↗ |
How do you run sandboxed WASM agents in Claude Code?
Run lightweight WebAssembly modules directly inside Claude, Cursor, or Codex agent sessions without leaving the IDE.
Who is it for?
Developers who need isolated WebAssembly agent sandboxes for untrusted code experiments inside Claude Code or Cursor.
Skip if: Developers who only need native shell execution without WASM isolation or portable cross-runtime agent packaging.
When should I use this skill?
The user needs to create, list, prompt, export, or terminate a sandboxed WASM agent for isolated code execution.
What you get
Running WASM agent instances, exported agent artifacts, and terminated sandbox sessions with isolated execution logs.
- Sandboxed WASM agent instances
- Exported agent artifacts
By the numbers
- Integrates 7 claude-flow wasm_agent MCP tools
- Exposes 4 primary CLI commands: create, list, prompt, terminate
Files
WASM Agent
Create sandboxed agents that run in WebAssembly for safe, isolated execution.
When to use
When you need to run untrusted code, experiment with agent configurations, or create portable agents that run anywhere WASM is supported.
Steps
1. Create agent — call mcp__claude-flow__wasm_agent_create with agent configuration 2. Send prompt — call mcp__claude-flow__wasm_agent_prompt to interact with the agent 3. Use tools — call mcp__claude-flow__wasm_agent_tool to give the agent access to specific tools 4. Manage files — call mcp__claude-flow__wasm_agent_files to read/write files in the sandbox 5. Export — call mcp__claude-flow__wasm_agent_export to package the agent for sharing 6. List agents — call mcp__claude-flow__wasm_agent_list to see all running WASM agents 7. Terminate — call mcp__claude-flow__wasm_agent_terminate to stop an agent
Benefits
- Full sandbox isolation — agents cannot access the host filesystem
- Portable — export and run on any WASM runtime
- Reproducible — same behavior across platforms
- Safe — no risk of system damage from agent actions
Related skills
Setup Matt Pocock SkillsScaffold the per-repo configuration that Matt Pocock’s engineering agent skills rely on so they understand the issue tracker, triage labels, and domain documentation la495k196k
Lark AppsConnect Claude, Cursor or custom agents directly to Lark (Feishu) for messaging, document automation, approval workflows and enterprise data access.402k
CavemanSlash token usage by roughly 75% while keeping every technical detail intact when working with Claude Code, Cursor or similar agents.391k94.7k
Lark Skill MakerQuickly turn any Lark/Feishu OpenAPI call or multi-step workflow into a reusable agent skill with its own SKILL.md.386k16k
Running Claude Code Via Litellm CopilotRun Claude Code at a fraction of the cost by routing requests through LiteLLM to the GitHub Copilot Chat API.270k72
Codex PetGenerate a complete Codex Pet spritesheet and metadata from one reference image without needing an OpenAI key or Codex Pro.246k10How it compares
Pick wasm-agent over raw Bash when isolation and portable WASM packaging matter more than full native filesystem access.
FAQ
What does wasm-agent sandbox protect against?
wasm-agent runs code inside WebAssembly sandboxes via claude-flow wasm_agent MCP tools, isolating untrusted execution from the host IDE environment. Developers experiment with agent configs and foreign code without granting full native shell access.
Which MCP tools power wasm-agent?
wasm-agent uses seven claude-flow MCP tools: wasm_agent_create, wasm_agent_list, wasm_agent_prompt, wasm_agent_tool, wasm_agent_files, wasm_agent_export, and wasm_agent_terminate for the full agent lifecycle.