AI agents use annotate to create or update resources in FiftyOne MCP Server — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your FiftyOne MCP Server environment.
Entering annotation mode for a sample is a Write operation as it involves creating or modifying annotation data on a dataset sample. The severity is medium since annotations are generally reversible but could affect dataset integrity if misused. Confidence is moderate because 'enter annotation mode' could simply be a UI state change (Read) rather than directly writing data.
From the tool's definition 'Enter annotation mode for a specific sample' - annotation implies creating or modifying labels/annotations on data
Documented attack patterns abuse exactly the kind of access annotate gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and FiftyOne MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for annotate:
{
"version": "1",
"default": "deny",
"tools": {
"annotate": {
"limits": [
{
"counter": "annotate_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}annotate stays usable, but capped — an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Enter annotation mode for a specific sample. It is categorised as a Write tool in the FiftyOne MCP Server MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the FiftyOne MCP Server MCP server in PolicyLayer and add a rule for annotate: 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 FiftyOne MCP Server. Nothing to install.
annotate 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 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. 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 is provided by the FiftyOne MCP Server MCP server (voxel51/fiftyone-mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from FiftyOne MCP Server, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
47 FiftyOne MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.