AI agents call recall_memory to retrieve information from Cockpit without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool queries historical classification patterns, correction history, and memory statistics from an accounting automation system. It is purely a retrieval/search operation with no side effects. While it accesses sensitive tax and transaction data (elevating severity to medium due to potential information disclosure risks in a financial context), the fundamental operation is Read-category.
From the tool's definition 過去の分類パターン・修正履歴を検索 (search past classification patterns and correction history); 過去パターンを参照 (reference past patterns); Memory stats の確認 (confirm memory statistics).
Attacks that exploit this kind of access
過去の分類パターン・修正履歴を検索。取引の摘要/取引先から過去パターンを参照し、分類結果の根拠を確認。Memory stats（pattern hit率等）の確認にも使用。. It is categorised as a Read tool in the Cockpit MCP Server, which means it retrieves data without modifying state.
Register the Cockpit MCP server in PolicyLayer and add a rule for recall_memory: 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 Cockpit. Nothing to install.
recall_memory 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_memory 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_memory. 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_memory is provided by the Cockpit MCP server (michielinksee/kansei-link-cockpit). 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 →