Stage 2. Returns cached external_facts plus a research directive for entities not yet known. Claude does web search at its layer and calls cache_research_facts.
AI agents call research_entities to retrieve information from Apple Calendar MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Even though research_entities only reads data, uncontrolled read access leaks sensitive information and racks up API costs — an agent caught in a retry loop can make thousands of calls a minute without anyone noticing.
Attacks that exploit this kind of access
Stage 2. Returns cached external_facts plus a research directive for entities not yet known. Claude does web search at its layer and calls cache_research_facts. It is categorised as a Read tool in the Apple Calendar MCP MCP Server, which means it retrieves data without modifying state.
Register the Apple Calendar MCP server in PolicyLayer and add a rule for research_entities: 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 Apple Calendar MCP. Nothing to install.
research_entities 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 research_entities 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 research_entities. 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.
research_entities is provided by the Apple Calendar MCP server (yongzhe-wang/yapping-apple-calendar-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.