Delete a memory entry. Writes a tombstone version, removes from all indexes.
AI agents call memory_delete to permanently remove resources in Memory — typically in cleanup and lifecycle workflows. It does its job in a single call, and there is no undo.
Although the operation writes a tombstone rather than physically erasing data, it removes the memory entry from all indexes and makes it inaccessible for retrieval, which is functionally destructive. The data is no longer queryable or retrievable by normal means. This is worse than Write (which is reversible) but the tombstone record prevents it from being classified as the most severe level.
From the tool's definition Tool description: 'Delete a memory entry. Writes a tombstone version, removes from all indexes.' The verb 'Delete' and phrase 'removes from all indexes' indicate irreversible removal of data.
Attacks that exploit this kind of access
Delete a memory entry. Writes a tombstone version, removes from all indexes. It is categorised as a Destructive tool in the Memory MCP Server, which means it can permanently delete or destroy data. Block by default and require explicit approval.
Register the Memory MCP server in PolicyLayer and add a rule for memory_delete: 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 Memory. Nothing to install.
memory_delete is a Destructive tool with critical risk. Critical-risk tools should be blocked by default and only enabled with explicit human approval.
Yes. Add a rate_limit block to the memory_delete 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_delete. 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_delete is provided by the Memory MCP server (joshdougall/memory-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 →