AI agents call fetch_metadata to retrieve information from Lyceum without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool only retrieves metadata information from online sources (Google Books) without modifying, deleting, or executing operations. It is purely informational and fits the Read category. Severity is low because metadata queries have minimal blast radius—they cannot damage data, trigger financial transactions, or execute arbitrary code.
From the tool's definition Tool name 'fetch_metadata' and description explicitly states it 'Search[es] online for book metadata' and 'Returns multiple results'. These are query/retrieval operations with no side effects on local or remote state.
Attacks that exploit this kind of access
Search online for book metadata by title, author, or ISBN. Returns multiple results from Google Books that the user can choose from. Use this when a book. It is categorised as a Read tool in the Lyceum MCP Server, which means it retrieves data without modifying state.
Register the Lyceum MCP server in PolicyLayer and add a rule for fetch_metadata: 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 Lyceum. Nothing to install.
fetch_metadata 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 fetch_metadata 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 fetch_metadata. 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.
fetch_metadata is provided by the Lyceum MCP server (matthewp/lyceum). 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 →