AI agents call upstream.mcp.attach to retrieve information from Dynamic without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The primary function is to query and retrieve tool metadata from another MCP server, which is a Read operation. However, the severity is elevated from 'low' to 'medium' because: (1) attachment to arbitrary external MCP servers could expose sensitive tool definitions or capabilities, (2) in a multi-tenant or restricted environment, this could be used to enumerate services an agent should not know about, and (3) the…
From the tool's definition Tool description states 'return its current tool list' — a retrieval operation with no side effects. However, 'attach to an existing MCP server' introduces network/connection complexity and potential information disclosure about external services.
Attacks that exploit this kind of access
Experimentally attach to an existing MCP server (stdio/http) and return its current tool list. It is categorised as a Read tool in the Dynamic MCP Server, which means it retrieves data without modifying state.
Register the Dynamic MCP server in PolicyLayer and add a rule for upstream.mcp.attach: 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 Dynamic. Nothing to install.
upstream.mcp.attach 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 upstream.mcp.attach 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 upstream.mcp.attach. 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.
upstream.mcp.attach is provided by the Dynamic MCP server (mcpland/dynamic-mcp). 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 →