graph_memory
AI agents call graph_memory to retrieve information from MemoryMesh without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool appears to retrieve or query relationships in an indexed knowledge graph without modifying data. No side effects, no data deletion, no code execution, and no financial operations are implied. The 'graph_' prefix suggests reading graph structures; sibling tools confirm this is a read-focused server. Low severity because it only accesses local, personal indexed data with no external side effects.
From the tool's definition Tool name 'graph_memory' on a semantic search and indexing server (MemoryMesh) with sibling tools like 'query_timeline', 'get_document', 'get_entity' that are all retrieval-oriented.
Attacks that exploit this kind of access
graph_memory. It is categorised as a Read tool in the MemoryMesh MCP Server, which means it retrieves data without modifying state.
Register the MemoryMesh MCP server in PolicyLayer and add a rule for graph_memory: 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 MemoryMesh. Nothing to install.
graph_memory 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 graph_memory 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 graph_memory. 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.
graph_memory is provided by the MemoryMesh MCP server (kilhubprojects/memory-mesh). 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 →