AI agents call find_superclasses to retrieve information from Orihime without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and queries class hierarchy information from the indexed knowledge graph without modifying any data or triggering side effects. It is a pure traversal operation on static code metadata, analogous to a read-only graph query. The low severity reflects that misuse poses minimal risk—returning incorrect class hierarchy information cannot execute code, delete data, or cause financial harm.
From the tool's definition Tool performs BFS traversal of the EXTENDS chain upward from a class with a maximum depth of 10. The description contains no mutation verbs (create, update, delete, execute) and explicitly indicates it 'walks' (traverses/queries) a code structure graph.
Attacks that exploit this kind of access
Walk the EXTENDS chain upward from the given class (BFS, max depth 10). It is categorised as a Read tool in the Orihime MCP Server, which means it retrieves data without modifying state.
Register the Orihime MCP server in PolicyLayer and add a rule for find_superclasses: 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 Orihime. Nothing to install.
find_superclasses 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 find_superclasses 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 find_superclasses. 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.
find_superclasses is provided by the Orihime MCP server (srinivasan-sundaresan95/orihime). 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 →