pull_file_from_sandbox
AI agents call pull_file_from_sandbox to retrieve information from MCP4Modal Sandbox without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The name implies reading/fetching a file from a cloud sandbox to the local context, which is a Read operation. Severity is medium because the sandbox may contain sensitive data or credentials. Confidence is lowered due to the empty description, but the sibling tool 'read_file_content_from_sandbox' and the naming pattern strongly suggest this is a file retrieval operation.
From the tool's definition Tool name: pull_file_from_sandbox — 'pull' and 'from_sandbox' suggest retrieving/downloading a file from a remote sandbox environment.
Attacks that exploit this kind of access
pull_file_from_sandbox. It is categorised as a Read tool in the MCP4Modal Sandbox MCP Server, which means it retrieves data without modifying state.
Register the MCP4Modal Sandbox MCP server in PolicyLayer and add a rule for pull_file_from_sandbox: 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 MCP4Modal Sandbox. Nothing to install.
pull_file_from_sandbox is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the pull_file_from_sandbox 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 pull_file_from_sandbox. 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.
pull_file_from_sandbox is provided by the MCP4Modal Sandbox MCP server (milkymap/mcp4modal_sandbox). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →