§4.3.1 — Retrieve the memories sharing a turn_id plus N entries
AI agents call recall_thread to retrieve information from N3MemoryCore Lite (Working Memory) without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool queries existing data from the working memory system without side effects. It is a retrieval operation similar to 'get' or 'fetch'. The sibling tools show the distinction: delete_* operations are destructive, save_* is write, and this recall operation is purely read.
From the tool's definition Tool retrieves memories by turn_id and returns N entries, explicitly using 'Retrieve' in the description. No modification, deletion, or execution occurs.
Attacks that exploit this kind of access
§4.3.1 — Retrieve the memories sharing a turn_id plus N entries. It is categorised as a Read tool in the N3MemoryCore Lite (Working Memory) MCP Server, which means it retrieves data without modifying state.
Register the N3MemoryCore Lite (Working Memory) MCP server in PolicyLayer and add a rule for recall_thread: 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 N3MemoryCore Lite (Working Memory). Nothing to install.
recall_thread 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 recall_thread 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 recall_thread. 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.
recall_thread is provided by the N3MemoryCore Lite (Working Memory) MCP server (NeuralNexusNote/n3mcmcp-lite). 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 →