// SCAN REPORT

Remember

33 tools. 15 can modify or destroy data without limits.

3 destructive tools with no built-in limits. Policy required.

Last updated: 30 Jun 2026

15 can modify or destroy data

18 read-only

33 tools total

Community server · catalogue entry verified 30/06/2026

How to control Remember ↓

TOOL MAP

What Remember exposes to your agents

Read (18) Write / Execute (12) Destructive / Financial (3)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous Remember tools

Destructive · Critical remember_confirm This tool executes pre-authorized destructive actions. Destructive · High remember_delete_memory Although soft-deletion preserves data at the database level for recovery, from a user and application perspective this tool irreversibly removes memories from active access. Destructive · High remember_delete_relationship This tool meets the Destructive category definition: it irreversibly deletes data that cannot be undone. Write · High remember_ghost_config This tool manages access control configuration for a persona/ghost system.

15 of Remember's 33 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Remember

PolicyLayer is an MCP gateway — it sits between your AI agents and Remember, and nothing reaches the server without passing your rules. These are the rules we recommend:

Deny destructive operations

{
  "remember_confirm": {
    "deny_if": [
      {
        "conditions": [],
        "on_deny": "Blocked by default. Requires approval."
      }
    ]
  }
}

Destructive tools should never be available to autonomous agents without human approval.

Rate limit write operations

{
  "remember_create_internal_memory": {
    "limits": [
      {
        "counter": "remember_create_internal_memory_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "remember_admin_collection_stats": {
    "limits": [
      {
        "counter": "remember_admin_collection_stats_per_minute",
        "window": "minute",
        "max": 60,
        "scope": "grant"
      }
    ]
  }
}

Controls API costs and prevents retry loops from exhausting upstream rate limits.

Create a free account and register Remember — nothing to install.
Add these rules — paste them, or build them visually. Tune the limits to your setup.
Point your MCP client (Claude, Cursor, anything) at your gateway URL.

ENFORCE POLICY ON REMEMBER →

Instant setup, no code required.

FULL CATALOGUE

All 33 Remember tools

DESTRUCTIVE 3 tools

Destructive remember_confirm Confirm and execute a pending action using the token. Works for any action that requires confirmation (publish Destructive remember_delete_memory Request to delete a memory. Requires confirmation via remember_confirm. ⚠️ IMPORTANT: This is a two-step pr Destructive remember_delete_relationship Delete a relationship from your collection. Automatically removes the relationship reference from all con

WRITE 12 tools

Write remember_create_internal_memory Create an internal memory (ghost observation or agent note) based on the current session context. In ghost Write remember_create_memory Create a new memory with optional template. Memories can store any type of information: notes, events, pe Write remember_create_relationship Create a relationship connecting 2 or more memories. Relationships describe how memories are connected wi Write remember_deny Deny a pending action. The request will be marked as denied and the token invalidated. Works for any action th Write remember_ghost_config Manage ghost/persona configuration. Controls who can interact with your ghost and at what trust level. Action Write remember_moderate Approve, reject, or remove a published memory. Requires moderator permissions (can_moderate). Actions: - appr Write remember_publish Publish a memory to one or more shared spaces and/or groups. The memory will be COPIED (not moved) from your p Write remember_request_set_trust_level Request a trust level change for a memory. Returns a confirmation token. Trust levels (1-5 integer scale): Write remember_revise Sync updated content from your source memory to all its published copies in spaces and groups. Generates a con Write remember_update_internal_memory Update an internal memory (ghost or agent). Only works on memories matching the current session Write remember_update_memory Update an existing memory with partial updates. Supports updating any field except id, user_id, doc_type, Write remember_update_relationship Update an existing relationship with partial updates. Supports updating relationship_type, observation, s

READ 18 tools

Read remember_admin_collection_stats [Admin] Get stats for a Weaviate collection — object count, property count, configuration. Requires admin a Read remember_admin_detect_weaviate_drift [Admin] Compare expected vs actual Weaviate schema properties per collection. Reports missing properties, e Read remember_admin_get_weaviate_schema [Admin] Inspect a Weaviate collection Read remember_admin_health [Admin] Deep health check — Weaviate and Firestore connectivity with latency. Returns overall status (healt Read remember_admin_inspect_memory [Admin] Fetch a raw memory object by UUID — all fields including internal metadata. Uses the Firestore inde Read remember_admin_inspect_user_api_tokens [Admin] Inspect a user Read remember_admin_inspect_user_escalation_records [Admin] Inspect a user Read remember_admin_inspect_user_ghost_configs [Admin] Inspect a user Read remember_admin_inspect_user_preferences [Admin] Inspect a user Read remember_admin_list_collections [Admin] List all Weaviate collections with type categorization (user, space, group). Optionally filter by p Read remember_admin_search_across_users [Admin] Search memories across multiple user tenants. Results include which user each memory belongs to. Read remember_find_similar Find memories similar to a given memory or text using vector similarity. Uses pure semantic similarity (v Read remember_get_core Get the ghost Read remember_query_internal_memory Query internal memories (ghost or agent) using natural language (pure semantic search). Automatically scope Read remember_search_internal_memory Search internal memories (ghost or agent) using hybrid semantic + keyword search. Automatically scoped to t Read remember_search_internal_memory_by Search internal memories (ghost or agent) using specialized sort/discovery modes. Automatically scoped to th Read remember_search_relationship Search relationships by observation text or relationship type. Uses semantic search on relationship obser Read remember_search_space_by Search published memories in shared spaces/groups using specialized sort/discovery modes. Modes: - byTime

EXPLORE

Related servers

Other MCP servers with similar tools — same risk classification, starter policies for each.

FAQ

Questions about Remember

Can an AI agent delete data through the Remember MCP server? +

Yes. The Remember server exposes 3 destructive tools including remember_confirm, remember_delete_memory, remember_delete_relationship. These permanently remove resources with no undo. PolicyLayer blocks destructive tools by default so they never reach the upstream server.

How do I prevent bulk modifications through Remember? +

The Remember server has 12 write tools including remember_create_internal_memory, remember_create_memory, remember_create_relationship. Set a rate limit in your policy -- for example, 10 calls per hour prevents an agent from making more than 10 modifications per hour. PolicyLayer enforces this at the gateway, before calls reach Remember.

How many tools does the Remember MCP server expose? +

33 tools across 3 categories: Destructive, Read, Write. 18 are read-only. 15 can modify, create, or delete data.

How do I enforce a policy on Remember? +

Register the Remember MCP server in PolicyLayer, apply the suggested rules above (adjust the limits to your use case), and point your AI client at the PolicyLayer proxy URL instead of the server directly. Your agents keep the same tools; PolicyLayer evaluates every call against policy before it executes. Nothing to install, live in minutes.

Enforce policy on every Remember tool call.

Deterministic rules across all 33 Remember tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON REMEMBER →

Instant setup, no code required.

33 Remember tools catalogued and risk-classified — across an index of 43,000+ MCP servers.