Refine a previous recall with a follow-up query. Use this when recall returned low-confidence results and hints suggested rephrasing. Combines the original query embedding with the refinement embedding (weighted: 0.4 original + 0.6 refinement) and excludes already-seen memory IDs. Args: original_...
AI agents call memory_recall_refine to retrieve information from Yantrikdb without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs semantic search refinement on a cognitive memory store. It reads and retrieves memory records based on query embeddings but does not create, modify, delete, or execute external operations. The exclusion of already-seen memory IDs is part of the search logic, not a destructive operation.
From the tool's definition The tool description explicitly states it "Refine a previous recall with a follow-up query" and "Combines the original query embedding with the refinement embedding" to retrieve results.
Documented attack patterns abuse exactly the kind of access memory_recall_refine gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Yantrikdb, and nothing reaches the server without passing your rules. This is the rule we recommend for memory_recall_refine:
{
"version": "1",
"default": "deny",
"tools": {
"memory_recall_refine": {}
}
}memory_recall_refine is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Refine a previous recall with a follow-up query. Use this when recall returned low-confidence results and hints suggested rephrasing. Combines the original query embedding with the refinement embedding (weighted: 0.4 original + 0.6 refinement) and excludes already-seen memory IDs. Args: original_query: The original search query text. refinement_text: The follow-up/clarifying query text. original_rids: Memory IDs from the first recall to exclude from results. top_k: Maximum number of results (default 10). namespace: Filter by namespace. domain: Filter by domain. source: Filter by source. Returns new results with confidence and hints. It is categorised as a Read tool in the Yantrikdb MCP Server, which means it retrieves data without modifying state.
Register the Yantrikdb MCP server in PolicyLayer and add a rule for memory_recall_refine: 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 Yantrikdb. Nothing to install.
memory_recall_refine 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_recall_refine 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_recall_refine. 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_recall_refine is provided by the Yantrikdb MCP server (yantrikos/yantrikdb-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 14 Yantrikdb tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
14 Yantrikdb tools catalogued and risk-classified — across an index of 42,500+ MCP servers.