AI agents use create_relations to create or update resources in Rawthink — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Rawthink environment.
This tool creates relations (edges/links) in a knowledge graph—a reversible write operation. It modifies the graph structure but does not delete or irreversibly destroy data. Severity is medium because uncontrolled relation creation could pollute the knowledge graph, create false associations, or degrade system utility, but the impact is bounded to the memory system and reversible via delete_relations.
From the tool's definition Tool name is 'create_relations' on a knowledge graph / persistent memory system. Sister tools include 'delete_relations', 'delete_entities', and 'delete_observations', establishing that this server manages persistent graph data.
Attacks that exploit this kind of access
create_relations. It is categorised as a Write tool in the Rawthink MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Rawthink MCP server in PolicyLayer and add a rule for create_relations: 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 Rawthink. Nothing to install.
create_relations 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 create_relations 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 create_relations. 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.
create_relations is provided by the Rawthink MCP server (ygtalp/rawthink-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 →