HUMAN-IN-THE-LOOP GATE. Commit a staged write.
AI agents use approve_write to create or update resources in Fhir Synthetic — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Fhir Synthetic environment.
This tool commits a previously proposed write operation (from propose_observation) after human approval. While the human-in-the-loop gate mitigates risk by requiring explicit approval, the tool itself performs an irreversible state change to the FHIR data system. The structured audit logging further reduces severity from high to medium.
From the tool's definition Tool name is 'approve_write' and description states 'Commit a staged write', indicating it finalizes and applies a write operation to the system.
Attacks that exploit this kind of access
HUMAN-IN-THE-LOOP GATE. Commit a staged write. It is categorised as a Write tool in the Fhir Synthetic MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Fhir Synthetic MCP server in PolicyLayer and add a rule for approve_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 Fhir Synthetic. Nothing to install.
approve_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 approve_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 approve_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.
approve_write is provided by the Fhir Synthetic MCP server (krishnakakani-github/fhir-synthetic-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 →