AI agents use annotate_cluster to create or update resources in Millimap — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Millimap environment.
Annotation is a reversible write operation that modifies metadata associated with clusters in the active MilliMap session. While the empty description reduces confidence slightly, the name and context of cluster annotation tools typically create or update labels/metadata rather than executing arbitrary code or deleting data.
From the tool's definition Tool name 'annotate_cluster' indicates modification of cluster annotations in the MilliMap analysis session. The server description states it can 'drive analyses' including 'annotations', and this tool name suggests creating or updating cluster annotations.
Attacks that exploit this kind of access
annotate_cluster. It is categorised as a Write tool in the Millimap MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Millimap MCP server in PolicyLayer and add a rule for annotate_cluster: 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 Millimap. Nothing to install.
annotate_cluster 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 annotate_cluster 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 annotate_cluster. 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.
annotate_cluster is provided by the Millimap MCP server (milliomics/millimap-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 →