AI agents use memory_rollback to create or update resources in Memory — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Memory environment.
This tool modifies a memory entry by reverting it to a prior state, but the action is recorded as a new version, making it reversible. It does not irreversibly destroy data — the current version still exists in history — so it falls under Write rather than Destructive. Misuse could corrupt agent memory state, warranting medium severity.
From the tool's definition Restore a memory entry to a previous version. The rollback itself is recorded as a new version.
Attacks that exploit this kind of access
Restore a memory entry to a previous version. The rollback itself is recorded as a new version. It is categorised as a Write tool in the Memory MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Memory MCP server in PolicyLayer and add a rule for memory_rollback: 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_rollback 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_rollback 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_rollback. 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_rollback 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 →