// SCAN REPORT

Claude Memory MCP

37 tools. 23 can modify or destroy data without limits.

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

Last updated: 3 Jul 2026

23 can modify or destroy data

14 read-only

37 tools total

Community server · catalogue entry verified 03/07/2026

How to control Claude Memory MCP ↓

TOOL MAP

What Claude Memory MCP exposes to your agents

Read (14) Write / Execute (21) Destructive / Financial (2)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous Claude Memory MCP tools

Destructive · High memory_delete Hard-delete operations are irreversible and permanently remove data. Destructive · High memory_delete_rule This tool deletes rules that enforce constraints within the memory system. Execute · Medium memory_reembed This tool triggers a bulk computational operation (re-embedding) across all active memories using the current model. Write · Medium memory_add_rule This tool creates or modifies configuration data (rules) persistently.

23 of Claude Memory MCP's 37 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Claude Memory MCP

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

Deny destructive operations

{
  "memory_delete": {
    "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

{
  "memory_add_rule": {
    "limits": [
      {
        "counter": "memory_add_rule_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "memory_check_update": {
    "limits": [
      {
        "counter": "memory_check_update_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 Claude Memory MCP — 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 CLAUDE MEMORY →

Instant setup, no code required.

FULL CATALOGUE

All 37 Claude Memory MCP tools

DESTRUCTIVE 2 tools

Destructive memory_delete Soft-delete (archive) or hard-delete a memory. Destructive memory_delete_rule Delete a rule by its id. Soft-deletes (archives) unless hard=True.

EXECUTE 1 tools

Execute memory_reembed Re-embed all active memories with the current model.

WRITE 20 tools

Write memory_add_rule Add a project rule. rule_type is 'mandatory' (always do) or 'forbidden' Write memory_add_rule_bulk Add one rule to many projects at once. Write memory_add_template_rule Add a rule to a template (by template name). rule_type is 'mandatory' or Write memory_apply_template Apply a template's rules/memories into a project (by template name). Write memory_attach_project Attach an existing project directory. Auto-activates on success. Write memory_create_template Create a reusable template - a named set of default rules/memories that Write memory_import Import memories from exported .md files. Write memory_import_claude_md memory_import_claude_md Write memory_import_rules Copy selected rules/memories from another project into this one. Use Write memory_init_project Initialize a new project namespace (creates DuckDB + registers it). Write memory_link_folder Bind a project to a source folder for git-synced memory. Write memory_make_portable Move the project's DB into the project directory for git sharing. Write memory_rename_project Rename a project (its display name) and optionally update its description. Write memory_session_end End a session and store its summary. Write memory_set_model Switch embedding model between 'english' and 'multilingual' presets. Write memory_store Store a new memory with auto-embedding, summary, entity extraction, and TTL. Write memory_sync Sync a portable DB after git pull. Auto-activates on success. Write memory_update Update an existing memory. Re-embeds if title/content changed. Write memory_update_rule Update an existing mandatory or forbidden rule by its id. Write memory_use Set the active project. Subsequent tools use it by default.

READ 14 tools

Read memory_check_update Check if a newer version of the Memory MCP server is available. Read memory_export Export all active memories to human-readable .md files. Read memory_get_rules Get all mandatory and forbidden rules (direct SQL, cached). Read memory_list List memories with filtering, sorting, and pagination. Read memory_list_projects List all registered projects. Read memory_list_templates List reusable rule/memory templates that can be applied to new projects. Read memory_load_from_folder memory_load_from_folder Read memory_model_info Current embedding model + available presets. Read memory_project_info Get detailed info for a project. Read memory_provenance Get the full audit trail for a memory. Read memory_recall Recall a specific memory by ID or exact title. Read memory_search Semantic search with composite relevance scoring. Read memory_session_start Start a session. Loads rules, last summary, sprint goals, recent decisions. Read memory_version Get the current version of the Memory MCP server and configuration.

EXPLORE

Related servers

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

FAQ

Questions about Claude Memory MCP

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

Yes. The Claude Memory MCP server exposes 2 destructive tools including memory_delete, memory_delete_rule. 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 Claude Memory MCP? +

The Claude Memory MCP server has 20 write tools including memory_add_rule, memory_add_rule_bulk, memory_add_template_rule. 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 Claude Memory MCP.

How many tools does the Claude Memory MCP server expose? +

37 tools across 4 categories: Destructive, Execute, Read, Write. 14 are read-only. 23 can modify, create, or delete data.

How do I enforce a policy on Claude Memory MCP? +

Register the Claude Memory 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 Claude Memory MCP tool call.

Deterministic rules across all 37 Claude Memory MCP tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON CLAUDE MEMORY →

Instant setup, no code required.

37 Claude Memory MCP tools catalogued and risk-classified — across an index of 43,000+ MCP servers.