// SCAN REPORT

Loom

31 tools. 20 can modify or destroy data without limits.

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

Last updated: 30 Jun 2026

20 can modify or destroy data

11 read-only

31 tools total

Community server · catalogue entry verified 30/06/2026

How to control Loom ↓

TOOL MAP

What Loom exposes to your agents

Read (11) Write / Execute (16) Destructive / Financial (4)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous Loom tools

Destructive · High forget This tool irreversibly deletes agent memories by reference or by category+title lookup. Destructive · High knowledge_purge This tool permanently removes archived knowledge and propagates deletions to dependent citations, making the action irreversible. Destructive · High memory_prune This tool permanently deletes data (expired memories) from persistent storage. Destructive · Medium memory_reject The tool explicitly deletes a staging row and discards a pending proposal.

20 of Loom's 31 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Loom

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

Deny destructive operations

{
  "forget": {
    "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

{
  "bootstrap": {
    "limits": [
      {
        "counter": "bootstrap_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "dossier": {
    "limits": [
      {
        "counter": "dossier_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 Loom — 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 LOOM →

Instant setup, no code required.

FULL CATALOGUE

All 31 Loom tools

DESTRUCTIVE 4 tools

Destructive forget Remove memories. Single deletion by ref or category+title. Destructive knowledge_purge Hard-delete one or more archived knowledge pages and cascade their citations. Destructive memory_prune Remove expired memories (TTL elapsed). Use dry_run to preview without deleting. Destructive memory_reject Discard a pending proposal without committing it. Deletes the staging row;

WRITE 16 tools

Write bootstrap Initialize a new loom identity from scratch. Generates IDENTITY.md, preferences.md, Write harness_init Scaffold a harness manifest at <contextDir>/harnesses/<name>.md from the template Write knowledge_archive Soft-retire a knowledge page: set its status to archived with an optional tombstone note. Write knowledge_merge Consolidate 2+ knowledge pages into one canonical page. Write knowledge_move Re-key or re-domain a knowledge page in place — same row, same uuid, citations and verification history preser Write knowledge_restore Restore a previously archived knowledge page back to active status. Write knowledge_supersede Mark one knowledge page as superseded by another, then archive the old page. Write knowledge_verify Stamp a knowledge page as verified WITHOUT touching its body — sets verified_at Write knowledge_write Upsert an entity page by slug. On an existing slug: body REPLACES by default Write memory_archive Soft-retire a memory: move it to the archive tier with a tombstone instead of Write memory_propose Stage a DRAFT memory in the capture-propose queue for later ratification. Write memory_ratify Ratify a pending proposal into a REAL memory. Loads the proposal, applies any Write memory_restore Restore a previously archived memory to the active set. Clears the archive Write remember Store an episodic memory that persists across sessions. Use this when you Write update Update an existing memory. Find by ref (returned from remember) or by Write update_identity Update your self-model or preferences with section-level precision.

READ 11 tools

Read dossier Load Art\ Read find_similar Surface memories semantically near an existing ref or free-form text. Read harness_describe Self-describe the CURRENTLY CONNECTED harness: write its manifest at Read identity Load the persistent identity for this agent. Returns the terminal creed Read knowledge_history Body-revision history for a knowledge page. Replace-writes snapshot the displaced Read knowledge_maintain Read-only health report for the knowledge store. Three branches: Read knowledge_recall Search the knowledge store with LIKE matching over title, body, and domain, Read memory_audit One-shot health report for the memory store: totals, category breakdown, Read memory_list Browse memories without semantic search. Lists memories with optional Read memory_proposals List all pending proposals in the capture-propose queue, newest first. Read recall Retrieve memories relevant to a query or topic. Returns matching memories

EXPLORE

Related servers

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

FAQ

Questions about Loom

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

Yes. The Loom server exposes 4 destructive tools including forget, knowledge_purge, memory_prune. 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 Loom? +

The Loom server has 16 write tools including bootstrap, harness_init, knowledge_archive. 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 Loom.

How many tools does the Loom MCP server expose? +

31 tools across 3 categories: Destructive, Read, Write. 11 are read-only. 20 can modify, create, or delete data.

How do I enforce a policy on Loom? +

Register the Loom 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 Loom tool call.

Deterministic rules across all 31 Loom tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON LOOM →

Instant setup, no code required.

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

// WHERE THIS COMES FROM

These policies come from Loom's registry record.

The record behind this page: verified identity, auth posture, risk grade, every tool classified, recommended policy — re-checked continuously.

Teams ship this data inside their own products. See what a licence covers →