AI agents use memory_link to create or update resources in Exocortex — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Exocortex environment.
The tool writes graph edges (links) between memory entries. Creating a link is a reversible write operation. Removing a link is also covered ('or remove'), which could be considered mildly destructive, but since links can be recreated, it remains reversible. The most severe applicable category is Write. Misuse could corrupt memory graph topology, affecting retrieval quality for the AI agent, hence medium severity.
From the tool's definition 'Create or remove a link between two memories' — creates or modifies relationships between existing memory nodes
Attacks that exploit this kind of access
Create or remove a link between two memories. Links enable graph-aware context retrieval — linked memories surface together during context loading. It is categorised as a Write tool in the Exocortex MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Exocortex MCP server in PolicyLayer and add a rule for memory_link: 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 Exocortex. Nothing to install.
memory_link 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 memory_link 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 memory_link. 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.
memory_link is provided by the Exocortex MCP server (shawnhack/exocortex). 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 →