Safely run untrusted commands for MCP clients inside a gVisor sandbox.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "io.github.IronSecCo/ironclaw" yet — see the docs or source repo.
Use ironclaw to run the following agent-generated shell command inside a gVisor sandbox, and return stdout, stderr, and the exit code: python script.py --input sample.json
The tool returns the command result from the sandbox, including output, errors, and exit status, without affecting the host environment.
Run this potentially risky install command in the ironclaw sandbox, observe what files it creates and what logs it outputs, then summarize the potential risks: curl -fsSL https://example.com/install.sh | bash
You get execution logs and a behavior summary from the isolated environment to help assess the script's safety.
Use ironclaw to reproduce the following build command in a sandbox and capture the cause of failure: npm ci && npm run build. Please provide full logs and possible fixes.
The output includes build logs, failure diagnosis, and suggested fixes for dependency or environment issues.
Run AI agents in secure microVM sandboxes with network and privacy guardrails.
Run AI agents in VM-isolated sandboxes on Mac for safer execution.
Run Git and GitHub CLI outside sandbox restrictions for smoother repository operations.
Safely run whitelisted shell commands with timeouts, pipelines, and audit logs.
Execute shell commands securely over MCP for controlled dev and ops workflows.
Securely orchestrate SSH commands, files, tunnels, and fleet operations across hosts.