Search historical memories. Direct results = vector DB; related = graph DB. Filter by categories, tags, observation_types, concepts (same dimensions as save_memory). Call BEFORE answering about preferences/history. time_filter: absolute dates (YYYY-MM-DD). entity_name: graph-focused search by fil...
AI agents call search_memory to retrieve information from Memoryx without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
| Parameter | Type | Required | Description |
|---|---|---|---|
tags | array | — | Optional. Filter by tags |
limit | number | — | Max results, default 5 |
query | string | Yes | Search intent: keywords or natural language question |
concepts | array | — | Optional. Filter by concepts (how-it-works/why-it-exists/what-changed/problem-solution/gotcha/pattern/trade-off) |
categories | array | — | Optional. Filter by category (semantic/episodic/procedural/emotional/reflective) |
entity_name | string | — | Optional. Expand graph from this entity (e.g. file path, class name, method); same convention as server extraction |
time_filter | object | — | Optional. start/end in YYYY-MM-DD or YYYY-MM-DD HH:mm:ss (absolute dates only) |
observation_types | array | — | Optional. Filter by observation_type (bugfix/feature/refactor/change/discovery/decision) |
Parameters from the server's own tool schema.
This tool retrieves and queries memory data using vector and graph database searches, filtered by various dimensions. It has no side effects—it only returns results and memory_ids for potential use with other tools. This is a standard Read operation with minimal risk, appropriate for an AI agent to call when checking historical context.
From the tool's definition Tool description explicitly states it "Search[es] historical memories" with "Direct results = vector DB; related = graph DB". Returns data without modifying it. No create, update, delete, or execute operations mentioned.
Risk signalsAccepts freeform code/query input (query)
Attacks that exploit this kind of access
Search historical memories. Direct results = vector DB; related = graph DB. Filter by categories, tags, observation_types, concepts (same dimensions as save_memory). Call BEFORE answering about preferences/history. time_filter: absolute dates (YYYY-MM-DD). entity_name: graph-focused search by file/class/method. Returned id = memory_id for delete_memory. It is categorised as a Read tool in the Memoryx MCP Server, which means it retrieves data without modifying state.
search_memory accepts 8 parameters: tags, limit, query, concepts, categories, entity_name, time_filter, observation_types. Required: query. The full parameter table on this page comes from the server's own tool schema.
Register the Memoryx MCP server in PolicyLayer and add a rule for search_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 Memoryx. Nothing to install.
search_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 search_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 search_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.
search_memory is provided by the Memoryx MCP server (@t0ken.ai/memoryx-mcp-server). 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 →