Delete a memory by ID.
AI agents call memory_forget to permanently remove resources in Tartarus — typically in cleanup and lifecycle workflows. It does its job in a single call, and there is no undo.
This tool permanently removes stored memories from the persistent storage system. Deletion is an irreversible operation with no undo mechanism—once a memory is deleted by ID, it cannot be recovered. While the blast radius is contained to the agent's own memory store (not system-wide or financial), the destructive nature of permanent data loss makes this Destructive rather than Write.
From the tool's definition Tool name is 'memory_forget' with description 'Delete a memory by ID.' The verb 'Delete' combined with 'by ID' indicates irreversible removal of data from the SQLite-backed memory store.
Attacks that exploit this kind of access
Delete a memory by ID. It is categorised as a Destructive tool in the Tartarus MCP Server, which means it can permanently delete or destroy data. Block by default and require explicit approval.
Register the Tartarus MCP server in PolicyLayer and add a rule for memory_forget: 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 Tartarus. Nothing to install.
memory_forget 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_forget 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_forget. 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_forget is provided by the Tartarus MCP server (saluca-labs/tartarus-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 →