Generate accessible preview URLs for web applications running in the Daytona workspace. Creates a secure tunnel to expose local ports externally without configuration. Validates if a server is actually running on the specified port and provides diagnostic information for troubleshooting. Supports...
AI agents invoke web_preview to trigger actions in Daytona MCP Python Interpreter. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
While the tool itself appears focused on tunneling rather than direct code execution, it operates within an execution-oriented server (Python interpreter with shell_exec capability) and performs non-trivial infrastructure operations (tunnel creation, port validation, external exposure). These actions can trigger side effects external to the workspace.
From the tool's definition The tool "Generate accessible preview URLs for web applications" and "Creates a secure tunnel to expose local ports externally" indicates it executes operations that establish external network connections and tunnel creation.
Documented attack patterns abuse exactly the kind of access web_preview gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Daytona MCP Python Interpreter, and nothing reaches the server without passing your rules. This is the rule we recommend for web_preview:
{
"version": "1",
"default": "deny",
"tools": {
"web_preview": {
"limits": [
{
"counter": "web_preview_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}web_preview stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Generate accessible preview URLs for web applications running in the Daytona workspace. Creates a secure tunnel to expose local ports externally without configuration. Validates if a server is actually running on the specified port and provides diagnostic information for troubleshooting. Supports custom descriptions and metadata for better organization of multiple services. It is categorised as a Execute tool in the Daytona MCP Python Interpreter MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Daytona MCP Python Interpreter MCP server in PolicyLayer and add a rule for web_preview: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Daytona MCP Python Interpreter. Nothing to install.
web_preview is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the web_preview rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for web_preview. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
web_preview is provided by the Daytona MCP Python Interpreter MCP server (nibzard/daytona-mcp-interpreter). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Daytona MCP Python Interpreter, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
5 Daytona MCP Python Interpreter tools catalogued and risk-classified — across an index of 43,000+ MCP servers.