memory.sweep
Retention: forget every fact the ledger learned of before an instant, in one space or all, skipping held facts, and remove each from every store. Without
This record as markdown: /tools/agent-custody/memory.sweep.md
What memory.sweep does on Agent Custody
AI agents call memory.sweep to permanently remove resources in Agent Custody, typically in cleanup and lifecycle workflows. It does its job in a single call, and there is no undo.
Why memory.sweep is rated Critical
An AI agent that decides to call memory.sweep doesn't hesitate, doesn't double-check, and doesn't stop at one. Whatever it removes from Agent Custody is gone. There is no undo for destructive operations.
Risk signalsBulk/mass operation — affects multiple targets
Attacks that exploit this kind of access
The rule that runs memory.sweep safely
PolicyLayer is an MCP gateway: it sits between your AI agents and Agent Custody, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For memory.sweep, this is the rule to start with:
memory.sweep is removed from the agent's tool list entirely, so the agent never calls it. The rest of the server keeps working.
The button opens the PolicyLayer dashboard: create your workspace, connect Agent Custody, apply this rule, and every memory.sweep call is checked against it from then on.
Questions about memory.sweep
Retention: forget every fact the ledger learned of before an instant, in one space or all, skipping held facts, and remove each from every store. Without. It is categorised as a Destructive tool in the Agent Custody MCP Server, which means it can permanently delete or destroy data. Block by default and require explicit approval.
Register the Agent Custody MCP server in PolicyLayer and add a rule for memory.sweep: 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 Agent Custody. Nothing to install.
memory.sweep 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.sweep 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.sweep. 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.sweep is provided by the Agent Custody MCP server (ch4r10t33r/agent-custody). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
More on Agent Custody, and thousands of servers like it.
Across the catalogue