AI agents call memory_inject to retrieve information from Mementos without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and queries memory data based on selection criteria. Despite the word 'inject' in the name, the description indicates it pulls memory context for use in system prompts rather than modifying or executing anything. The action is read-only retrieval with no side effects. Blast radius is minimal—misuse would surface stale or irrelevant memory, not corrupt systems or trigger unintended operations.
From the tool's definition Tool description states 'Get memory context' and 'Selects by scope, importance, recency' — purely a retrieval operation with no modification, deletion, or execution of external actions.
Attacks that exploit this kind of access
Get memory context for system prompt injection. Selects by scope, importance, recency. Use strategy=. It is categorised as a Read tool in the Mementos MCP Server, which means it retrieves data without modifying state.
Register the Mementos MCP server in PolicyLayer and add a rule for memory_inject: 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 Mementos. Nothing to install.
memory_inject 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 memory_inject 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_inject. 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_inject is provided by the Mementos MCP server (@hasna/mementos). 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 →