Classify text into categories using Apple
AI agents call fm_classify to retrieve information from Pypi:apple Fm without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Text classification is a read/inference operation: it takes text as input and returns a categorical label. No data is written, deleted, or executed. The tool runs an on-device LLM inference call, which is analogous to a query. Severity is low because misuse is limited to receiving potentially mislabeled classifications; no external state is modified.
From the tool's definition "Classify text into categories" — the tool reads/analyzes input text and returns a classification result with no side effects.
Attacks that exploit this kind of access
Classify text into categories using Apple. It is categorised as a Read tool in the Pypi:apple Fm MCP Server, which means it retrieves data without modifying state.
Register the Pypi:apple Fm MCP server in PolicyLayer and add a rule for fm_classify: 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 Pypi:apple Fm. Nothing to install.
fm_classify 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 fm_classify 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 fm_classify. 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.
fm_classify is provided by the Pypi:apple Fm MCP server (pypi:apple-fm-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 →