AI agents use relay_write to create or update resources in Mcp Ad4m — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Mcp Ad4m environment.
This tool creates and modifies relay messages in a shared semantic memory system accessible across multiple sessions. It is reversible (messages can be cleared or overwritten) and does not permanently delete data or move financial resources, placing it in the Write category rather than Destructive or Financial.
From the tool's definition Tool name includes 'write'; description states it 'Write[s] a cross-terminal relay message to AD4M', explicitly indicating message creation/transmission to shared state.
Attacks that exploit this kind of access
Write a cross-terminal relay message to AD4M. Both Claude Code sessions connect to the same AD4M executor, so this enables real-time state sharing between terminals. It is categorised as a Write tool in the Mcp Ad4m MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Mcp Ad4m MCP server in PolicyLayer and add a rule for relay_write: 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 Mcp Ad4m. Nothing to install.
relay_write is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the relay_write 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 relay_write. 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.
relay_write is provided by the Mcp Ad4m MCP server (thefranceway/mcp-ad4m). 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 →