Skip to main content
Glama
Automox MCP Server

Automox MCP Server

Automox MCP Server

CI Security Scans Publish Release PyPI version

The official MCP server for Automox. Talk to your Automox console using natural language — this MCP server connects AI assistants like Claude to your Automox environment so you can manage devices, check compliance, run policies, and more, just by asking.

You:   "Are we ready for Patch Tuesday?"
Claude: Here's your readiness summary — 3 devices need patches,
        2 approvals are pending, and your patch policies run tonight at 2 AM...
IMPORTANT

For bug reports or feature requests, usehelp.automox.com or your typical escalation paths.

CAUTION

AI assistants can make mistakes. Responses produced by the MCP server may be incorrect or incomplete. If you see this happening consistently, please let us know.

Table of Contents

Related MCP server: Okta MCP Server

What's New in 3.0

Automox MCP Server 3.0 adds a centrally hosted option to replace the legacy self-hosted solution. The hosted server supports all the functionality you're already using, plus one new capability.

  • Hosted server: Automox now runs and maintains a version of this same server for you. Nothing to install, connect your AI client to https://console.automox.com/api/mcp with your existing API key. See Self-Hosted vs. Hosted below.

  • Policy Catalog templates (hosted only): The hosted service can search Automox's library of best-practice policy templates and create a policy directly from one, so you have a starting point without building a policy from scratch. This is separate from asking about the policies already deployed in your org, this is Automox's own recommended template library.

  • No capability loss: The hosted service exposes the same tool coverage you already have through this repository.

  • No forced migration: This self-hosted server and the Claude Desktop extension keep working exactly as they do today. Moving to the hosted service is currently optional (but recommended), see Migrating to the Hosted Server.

  • Auth is unchanged for now: Both versions use the same Automox API key. SSO is planned for a future release.

Self-Hosted vs. Hosted

Both run the same open-source server code. The difference is who installs and runs it, and one capability that's currently hosted-only.

Self-hosted (this repository)

Hosted (MCP Server 3.0)

Install

You install and keep it updated (PyPI, uvx, the Claude Desktop extension, or via Claude's Connectors Directory)

Nothing to install, Automox runs it

Where it runs

Your machine

Automox's multi-tenant service

Auth

Your Automox API key

Your Automox API key (same model, for now)

Claude Desktop

Works today via the one-click extension, installable from GitHub Releases or Claude's Connectors Directory

Not yet, Claude Desktop's built-in custom connector requires OAuth, which the hosted service doesn't support yet

Other clients (Claude Code, Cursor, MCP Inspector, etc.)

Works via your local config

Works today over HTTP with a Bearer token header

Good fit if

You want to run your own infrastructure, or you use Claude Desktop today

You want zero maintenance and use a client that supports custom HTTP headers

For self-hosted setup, see Quick Start below. For the hosted server, see Migrating to the Hosted Server, which also covers connecting fresh with no prior install.

Quick Start

1. Get your Automox credentials

You need three values from the Automox Console:

Value

Where to find it

API Key

Use an org-scoped key — zone Settings > Secrets & Keys > Add API Key (docs); see key types below

Account UUID

Settings > Secrets & Keys (shown on the page)

Org ID

The numeric ID in the URL when viewing your organization

Automox has two API key types, and the difference matters here:

Org-scoped key (recommended)

Global / account key

Scope

One organization — the zone it was created in

Every org in the account; inherits the key owner's role per org

Created at

Zone Settings > Secrets & Keys

Account Global Access Management > Keys (Full Administrator)

Tool coverage

All tools (verified: works immediately on the search family)

Unreliable on the Advanced Device Search family — advanced_device_search, device_search_typeahead, saved-search create/read/update/delete, list_searches_for_device, get_device_assignments — observed returning 403 in most orgs even for full administrators, while working in others; the upstream authorization behavior is inconsistent and the mechanism is unconfirmed

Symptom: 403 on the search tools while reads work everywhere else usually means the key, not your permissions — switch to an org-scoped key for the target org. API Key and Account UUID are always required. Org ID is recommended but optional — some tools that don't require org context will work without it.

2. Create a .env file

AUTOMOX_API_KEY=your-api-key
AUTOMOX_ACCOUNT_UUID=your-account-uuid
AUTOMOX_ORG_ID=your-org-id

3. Connect to your AI assistant

Claude Desktop (recommended) — one-click MCPB install:

New: Automox MCP is now also listed directly in Claude's Connectors Directory. Open Claude Desktop, go to Settings > Connectors, and search "Automox" to connect without leaving the app. The manual steps below still work too, nothing about them has changed.

  1. Download the latest automox-mcp-<version>.mcpb from the GitHub Releases page.

  2. Open Claude Desktop → Settings → Extensions.

  3. Drag the .mcpb file into the Extensions window.

  4. Paste your API key, Account UUID, and (optionally) Org ID into the prompts.

No .env file, no terminal — credentials are stored in Claude Desktop's secure config. The bundle pulls the matching automox-mcp release from PyPI on first run.

Claude Code (CLI):

claude mcp add automox-mcp uvx -- --env-file /path/to/.env automox-mcp

Cursor / any other MCP client — add to your MCP config:

{
  "mcpServers": {
    "automox-mcp": {
      "command": "uvx",
      "args": ["--env-file", "/path/to/.env", "automox-mcp"]
    }
  }
}

That's it. Start asking questions.

What Can I Ask?

The server exposes 130+ tools across devices, policies, patches, groups, webhooks, worklets, vulnerability sync, maintenance windows, and more. You don't need to know the tool names — just describe what you want:

Ask this

What happens

"Are we ready for Patch Tuesday?"

Checks pending patches, approvals, and policy schedules

"What is our compliance posture?"

Returns compliance rates, non-compliant devices, and health breakdown

"Give me the full profile for the Caldera server"

Combines device details, inventory, packages, and policy status

"What devices need attention?"

Surfaces devices flagged for immediate action

"Reboot the device 'Testing box'"

Searches for the device and issues a reboot command

"Create a patch policy for Firefox targeting the 'MCP testing' group"

Creates the policy with sensible defaults

"What did Mark Hansen do in Automox last week?"

Queries the audit trail across the date range

"Find all Windows devices not seen in 30 days"

Uses advanced device search with structured queries

"Show me vulnerability remediation status"

Lists action sets with issues, solutions, and progress

"Search the worklet catalog for USB security"

Browses community worklets with evaluation/remediation code

For the full list of tools, parameters, and MCP resources, see the Tool Reference.

Tip: You can also ask the server itself — the discover_capabilities tool returns all available tools organized by domain.

Configuration

Environment Variables

Applies only to self-hosted 2.x servers. Not applicable to the hosted 3.0+ server.

Variable

Required

Default

Description

AUTOMOX_API_KEY

Yes

—

Automox API key (org-scoped recommended — see key types)

AUTOMOX_ACCOUNT_UUID

Yes

—

Account UUID from Secrets & Keys

AUTOMOX_ORG_ID

Recommended

—

Numeric organization ID (required by most tools)

AUTOMOX_MCP_READ_ONLY

No

false

Disable all write operations (85 of 133 tools remain)

AUTOMOX_MCP_ALLOW_APPLY_REMEDIATION_ACTIONS

No

false

Opt in to the apply_remediation_actions tool, which patches/runs worklets on endpoints immediately. Off by default even in write mode.

AUTOMOX_MCP_ALLOW_SPLASHTOP_BULK_INSTALL_UNINSTALL

No

false

Opt in to the splashtop_bulk_install_uninstall tool, which installs/uninstalls the Splashtop client across an entire server group in one call. Off by default even in write mode.

AUTOMOX_MCP_ALLOW_DELETE_DEVICE

No

false

Opt in to the delete_device tool, which permanently deletes a device record and its history (DELETE /servers/{id}). Irreversible and not reconstructable through the MCP. Off by default even in write mode.

AUTOMOX_MCP_ALLOW_UPLOAD_POLICY_FILE

No

false

Opt in to the upload_policy_file tool, which uploads a local installer file to a Required Software policy. Reads from the local filesystem, so it also requires AUTOMOX_MCP_UPLOAD_ALLOWED_DIRS and only works on the stdio (local) transport. Off by default even in write mode.

AUTOMOX_MCP_UPLOAD_ALLOWED_DIRS

No

—

Comma-separated absolute directories upload_policy_file may read installers from. Required for that tool to register; paths are canonicalized and must resolve inside an allowed dir.

AUTOMOX_MCP_UPLOAD_MAX_BYTES

No

10737418240

Max installer size for upload_policy_file (default 10 GB, Automox's ceiling).

AUTOMOX_MCP_UPLOAD_TIMEOUT_SECONDS

No

3600

Upload read/write timeout for upload_policy_file (large installers need more than the default request timeout).

AUTOMOX_MCP_MODULES

No

all

Comma-separated list of modules to load (see below)

AUTOMOX_MCP_TOKEN_BUDGET

No

4000

Max estimated tokens per response before truncation

AUTOMOX_MCP_SANITIZE_RESPONSES

No

true

Sanitize API data to mitigate prompt injection

AUTOMOX_MCP_TOOL_PREFIX

No

—

Prefix all tool names (e.g., automox) to prevent cross-server collisions

AUTOMOX_MCP_LOG_FORMAT

No

text

Log format: text or json (structured JSON for SIEM integration)

AUTOMOX_MCP_TRANSPORT

No

stdio

Transport: stdio, http, sse, or streamable-http

AUTOMOX_MCP_HOST

No

127.0.0.1

Bind address for HTTP/SSE

AUTOMOX_MCP_PORT

No

8000

Bind port for HTTP/SSE

AUTOMOX_MCP_API_KEYS

No

—

Comma-separated MCP endpoint API keys for HTTP/SSE Bearer-token auth (e.g., key1,label:key2)

AUTOMOX_MCP_API_KEY_FILE

No

—

Path to a file containing MCP endpoint API keys (one per line)

AUTOMOX_MCP_OAUTH_ISSUER

No

—

OIDC issuer URL for JWT auth (e.g., https://auth.example.com/realms/main)

AUTOMOX_MCP_OAUTH_JWKS_URI

No

—

JWKS endpoint for JWT key rotation (auto-derived from issuer if omitted)

AUTOMOX_MCP_OAUTH_AUDIENCE

When JWT auth

—

Expected JWT audience claim (prevents token passthrough); required when AUTOMOX_MCP_OAUTH_ISSUER is set

AUTOMOX_MCP_OAUTH_SERVER_URL

No

—

Canonical server URL; enables RFC 9728 Protected Resource Metadata

AUTOMOX_MCP_OAUTH_SCOPES

No

—

Comma-separated required OAuth scopes

AUTOMOX_MCP_ALLOWED_ORIGINS

No

—

Extra allowed Origin headers for DNS rebinding protection (comma-separated)

AUTOMOX_MCP_ALLOWED_HOSTS

No

—

Extra allowed Host headers for DNS rebinding protection (comma-separated)

AUTOMOX_MCP_DNS_REBINDING_PROTECTION

No

true

Set to false to disable DNS rebinding protection (not recommended)

AUTOMOX_MCP_ALLOW_REMOTE_BIND

No

false

Allow binding to non-loopback addresses (required for 0.0.0.0 or external IPs)

Read-Only Mode

AUTOMOX_MCP_READ_ONLY=true

Disables all write operations. Only read-only tools are registered (85 of 133). Useful for auditing and monitoring.

Modular Loading

Load only the tool modules you need:

AUTOMOX_MCP_MODULES=devices,policies

Available modules: audit, audit_v2, devices, device_search, policies, policy_history, users, groups, events, reports, packages, webhooks, worklets, data_extracts, vuln_sync, compound, policy_windows

Both settings can be combined:

AUTOMOX_MCP_READ_ONLY=true
AUTOMOX_MCP_MODULES=devices,policies

HTTP Transport

For non-stdio deployments:

uvx --env-file .env automox-mcp --transport http --host 127.0.0.1 --port 8000

Endpoint Authentication

When deploying over HTTP or SSE, you can require authentication on the MCP endpoint (separate from the Automox API key). Two strategies are supported:

Static API keys (simple):

automox-mcp --generate-key                         # generate a key
export AUTOMOX_MCP_API_KEYS="amx_mcp_a1b2c3..."    # or use a key file

OAuth 2.1 / JWT (enterprise IdP integration):

export AUTOMOX_MCP_OAUTH_ISSUER="https://auth.example.com/realms/main"
export AUTOMOX_MCP_OAUTH_AUDIENCE="https://mcp.example.com"
export AUTOMOX_MCP_OAUTH_SERVER_URL="https://mcp.example.com"  # enables RFC 9728 metadata

Clients must include Authorization: Bearer <token> on every request. Unauthenticated requests receive 401 Unauthorized with proper WWW-Authenticate headers. No effect on stdio transport.

Security

The Automox MCP server is designed for enterprise deployment with defense-in-depth security controls.

Highlights:

  • Read-only mode (AUTOMOX_MCP_READ_ONLY) disables all 48 write tools

  • Module filtering (AUTOMOX_MCP_MODULES) for least-privilege tool loading

  • Correlation IDs on every tool call, forwarded to Automox API as X-Correlation-ID

  • Rate limiting (30 calls/60s) with token budget estimation and auto-truncation

  • API key isolation — stored as private attribute with per-request auth injection (no header storage)

  • Generic error responses — no internal paths, connection strings, or API keys in error output

  • Prompt injection mitigation — API response sanitization with Unicode normalization, homoglyph defense, HTML tag/script stripping, and reference-style markdown stripping

  • Webhook secret handling — secrets stripped from idempotency cache after creation

  • Structured JSON logging (AUTOMOX_MCP_LOG_FORMAT=json) for SIEM integration

  • Tool name prefixing (AUTOMOX_MCP_TOOL_PREFIX) to prevent cross-server collisions

  • Sigstore-signed releases with CycloneDX SBOM

  • SSRF prevention — webhook URLs validated against private/loopback IPs and cloud metadata endpoints

  • MCP endpoint authentication — static API keys or OAuth 2.1/JWT with audience binding and RFC 9728 Protected Resource Metadata

  • DNS rebinding protection — Origin and Host header validation on all HTTP/SSE connections per the MCP transport spec

  • Security response headers — X-Content-Type-Options, X-Frame-Options, CSP, Cache-Control: no-store, Strict-Transport-Security on all HTTP responses

  • Authentication rate limiting — blocks IPs after repeated auth failures to mitigate brute-force attacks

  • Remote bind protection — non-loopback HTTP/SSE binding requires explicit --allow-remote-bind opt-in

  • MCP Tool Annotations on all 130+ tools — readOnlyHint, destructiveHint, idempotentHint, and openWorldHint per the MCP Protocol specification, enabling client-side confirmation dialogs and safety guardrails

  • Interactive MCP Apps (io.modelcontextprotocol/ui) — inline review/approval surfaces for consequential flows: compliance triage, patch approval, policy blast-radius, remediation apply, and RBAC access certification. Apps-capable hosts render them inline; other hosts degrade gracefully to the structured tool output. Write-flow Apps drive the existing gated tools through the host's confirmation — no new tools, no new gates — and ship under the host's deny-all CSP (self-contained, no external/CDN loads)

  • 61 security hardening items (V-001 through V-182, S-001 through S-006) documented in CHANGELOG and SECURITY.md

Capability model. The server wraps 100% of the published Automox Console API and Webhooks API, with a single deliberate exception — secret-exposing endpoints are never wrapped (API-key decrypt, password-setting). Every destructive operation is either ask-first (host confirmation) or gated behind a default-off env flag. Concretely, three categorical rules:

  • Secrets are never handled — the server never returns secret material and never lets the model set it. Credentials enter only via environment/config; decrypt endpoints are not wrapped, password-setting is excluded, and secret fields are redacted from every projection. This is the only intentional omission.

  • Destructive operations are two-tier. Single-target, recoverable actions are ask-first (destructiveHint: true, surfaced as a host confirmation dialog, disabled entirely by read-only mode). Operations where per-call confirmation can't protect you — fleet-scale, self-lockout, or arbitrary model-authored code execution — are gated behind explicit, default-off env flags (AUTOMOX_MCP_ALLOW_APPLY_REMEDIATION_ACTIONS, AUTOMOX_MCP_ALLOW_SPLASHTOP_BULK_INSTALL_UNINSTALL, AUTOMOX_MCP_ALLOW_DELETE_DEVICE). Device deletion is gated, not omitted.

The full coverage map, the gating principle, and every intentional omission are documented in API Coverage & Intentional Omissions.

For vulnerability reporting and the full threat model, see SECURITY.md. For deployment hardening (containers, Kubernetes, MCP gateways, TLS, authentication), see the Deployment Security Guide. Security posture is benchmarked against the Wiz MCP Security Best Practices cheat sheet.

Note: For network-accessible deployments, enable endpoint authentication (static keys via AUTOMOX_MCP_API_KEYS or JWT via AUTOMOX_MCP_OAUTH_ISSUER) and/or place the server behind an MCP gateway or authenticating reverse proxy. TLS termination is the deployer's responsibility.

Privacy Policy

The Automox MCP server acts as a stateless proxy between your AI assistant and the Automox API.

Data collection: The server does not collect, store, or transmit any user data beyond what is required to fulfill API requests to the Automox platform. API credentials are read from environment variables at startup and used solely for authenticating requests to the Automox API.

Data usage: All data retrieved from the Automox API is returned directly to the AI assistant that initiated the request. The server performs response sanitization (Unicode normalization, HTML stripping) for prompt injection defense, but does not analyze, aggregate, or repurpose API data for any other purpose.

Third-party sharing: The server does not share data with any third parties. It communicates exclusively with the Automox API (console.automox.com) using the credentials you provide. No telemetry, analytics, or usage data is sent to the server authors or any other service.

Data retention: The server retains no persistent data between sessions. In-memory caches (idempotency keys, rate-limit counters) are cleared when the process exits. Structured logs, when enabled, are written to stderr and are the deployer's responsibility to manage and retain.

See the Automox MCP Server Privacy Policy for the full privacy policy (mirrored in PRIVACY.md).

Alternative Installation

The Quick Start above uses uvx which requires no installation. If you prefer a persistent install:

# Using uv
uv tool install automox-mcp

# Using pip
pip install automox-mcp

Then set the environment variables in your shell and run automox-mcp.

Updating

If you already have the server installed, update to the latest version:

# uvx (Quick Start method) — force a cache refresh
uvx --refresh automox-mcp

# uv tool install
uv tool upgrade automox-mcp

# pip
pip install --upgrade automox-mcp

Note: uvx automatically refreshes its cache roughly every 7 days, so most users will pick up new releases without action. Run uvx --refresh to get the latest immediately.

Migrating to the Hosted Server

Automox MCP Server 3.0 (hosted) runs the same tools as this repository, just without a local install to maintain. Migrating is two steps: remove your local server, then connect to the hosted one. There's no in-place upgrade and nothing to convert, no config file to move and no data to port. If you've never run the Automox MCP server before, skip straight to Step 2, there's no package to install.

Step 1: Remove the local server

Claude Code:

claude mcp remove automox

Cursor or other config-based clients: remove the automox-mcp entry from your MCP config file.

Claude Desktop / the MCPB extension: open Claude Desktop and go to Settings → Extensions. Find Automox and either toggle it off (you can turn it back on later) or choose Uninstall to remove it completely. Restart Claude Desktop afterward so the old tools no longer appear. Before you remove it, note that the hosted service can't be used from Claude Desktop yet. Claude Desktop's custom connector requires OAuth, which the hosted service doesn't support yet. If Claude Desktop is where you use Automox, keep the extension installed and only remove it when you're moving to a client that can connect to the hosted server, such as Claude Code or Cursor.

Step 2: Connect to the hosted server

Claude Code:

claude mcp add --transport http automox \
  https://console.automox.com/api/mcp \
  --header "Authorization: Bearer YOUR_AUTOMOX_API_KEY"

Open a new session and ask something like, "What's our compliance posture?"

MCP Inspector (for evaluation or debugging):

npx @modelcontextprotocol/inspector

In the UI: Transport Type = Streamable HTTP, URL = https://console.automox.com/api/mcp, Authentication → Bearer Token = your API key (paste the raw key). Click Connect, then Tools → List Tools.

Other MCP clients (Cursor and similar), any client that supports streamable HTTP with custom headers:

URL:      https://console.automox.com/api/mcp
Header:   Authorization: Bearer YOUR_AUTOMOX_API_KEY

Your permissions in Automox apply exactly as they do in the console and API. The MCP acts as you. Your self-hosted installation will keep working during a transition period — there's no rush to switch.

Troubleshooting

Symptom

Cause / fix

Adding the hosted server fails in Claude Desktop's "custom connector" screen

Claude Desktop's built-in custom connector only supports signing in with OAuth. The hosted service doesn't have OAuth yet, it only accepts an API key, so pasting the hosted URL in there will fail. This is a different thing from the Automox MCP desktop extension: install that instead (one-click .mcpb install, also in Claude's Connectors Directory). It runs the server locally on your machine and connects using the same API key you already have, so you still get full functionality in Claude Desktop, just through the extension rather than the custom connector screen.

401 Unauthorized

The Authorization: Bearer header is missing or the key is wrong. Send the key exactly as Bearer <key>.

"OAuth Authentication Failed" or an error about invalid JSON

Your client fell back to OAuth discovery after a 401. Fix the Bearer header rather than trying a custom header name.

403 Forbidden on some tools

Check your key's scope. Use an org-scoped key. Some user-management features require admin scopes.

Answers look off

AI assistants can make mistakes. Verify important results in the Automox Console before acting on them.

Frequently Asked Questions

Does my self-hosted 2.x server stop working? No. Your existing self-hosted installation continues to work, there's no change to it or to how you get it. The hosted service is simply a new option if you'd like to use it, on your own timeline, with no requirement to switch.

Do I lose any capabilities by moving to the hosted service? No. The hosted service runs the same tools you already have access to today.

Will I need to switch to SSO? Not yet. The hosted service uses the same API key model you use today. SSO and per-user authentication are planned for a future release, and we'll provide clear migration guidance when that happens.

I use Claude Desktop. What should I do? Keep using the Automox MCP desktop extension for now. Claude Desktop's native connector option requires OAuth, which the hosted service doesn't support yet.

Need help? Reach out through help.automox.com.

Development

git clone https://github.com/AutomoxCommunity/automox-mcp.git
cd automox-mcp
uv python install
uv sync --python 3.13 --dev

Testing

Interactive debugging with MCP Inspector:

fastmcp dev

Run unit tests:

uv run --python 3.13 --dev pytest

Run production smoke tests (requires Automox credentials):

uv run python tests/smoke_production.py

MCP Scanner

Static analysis with Cisco's MCP Scanner:

mcp-scanner \
  --analyzers yara \
  --format summary \
  stdio \
  --stdio-command uv \
  --stdio-arg run \
  --stdio-arg automox-mcp \
  --stdio-env AUTOMOX_API_KEY=test-api-key \
  --stdio-env AUTOMOX_ACCOUNT_UUID=test-account \
  --stdio-env AUTOMOX_ORG_ID=1 \
  --stdio-env AUTOMOX_MCP_SKIP_DOTENV=1

Versioning

Follows Semantic Versioning. Update pyproject.toml, commit, tag (e.g., v0.1.0), and push — the release workflow publishes to PyPI automatically.

License

MIT License. See LICENSE.

Support

The official Automox MCP server. For questions, bugs, or feature requests use help.automox.com.

To report a security vulnerability, see SECURITY.md — please do not open a public issue.

Available Tools

122 tools
apply_policy_changesApply Policy ChangesC
Destructive

Create or update Automox policies with automatic format correction.

Inputs: operations (required), preview, request_id. Returns: Structured output of the ``apply_policy_changes`` tool.

audit_events_ocsfAudit Events OCSFA
Read-onlyIdempotent

Query OCSF-formatted audit events from the Automox Audit Service v2. Filter by date (required) and event type name. The category_name filter is applied client-side against the event type_name prefix: the upstream events do NOT carry a category_name field, only an integer category_uid that maps 1:N across categories, so category filtering matches the type_name label prefix (authentication/entity_management/web_resource_activity are live-verified prefixes; account_change/user_access are spec-derived and unverified live). An unmappable category token leaves results unfiltered and sets metadata.applied_filters.category_name_matched=false (so an empty result is never mistaken for 'no activity'); metadata.events_before_filter reports the unfiltered count. category_uid/type_uid/class_uid/activity_id are raw OCSF taxonomy integers with no decode table in the upstream spec — prefer the human-readable sibling strings type_name and activity. Uses cursor-based pagination for large result sets. Event time is an ISO 8601 UTC string (converted from the upstream epoch-seconds value). The date parameter selects events by event date; the timezone of that date boundary is not stated by the upstream spec (unverified). severity/status labels follow the OCSF scales (severity: informational/low/medium/high/critical/fatal; status: success/failure/other) and are filled from severity_id/status_id when the upstream omits the string. Permissions: as of 2025-10-27 the upstream endpoint requires the API key to have BOTH organization:manage and users:read scopes; keys missing either scope return 403.

Inputs: date (required), limit, cursor, type_name, +2 more.

batch_update_devicesBatch Update DevicesA
Destructive

Apply bulk attribute actions to many devices at once (up to 500). Currently supports tag apply/remove via actions like {'attribute': 'tags', 'action': 'apply', 'value': ['env:prod']}.

Inputs: actions (required), devices (required), request_id. Returns: object

check_group_exclusion_statusCheck Group Exclusion StatusA
Read-onlyIdempotent

Check whether one or more server groups are currently within an active exclusion window. Returns a per-group boolean status.

Inputs: group_uuids (required), org_uuid. Returns: object

check_window_activeCheck Window ActiveA
Read-onlyIdempotent

Check whether a specific maintenance window is currently active. A window is active when its status is 'active', it has at least one group, and the current time falls within an exclusion period.

Inputs: window_uuid (required), org_uuid. Returns: object

clone_policyClone PolicyA
Destructive

Clone an existing Automox policy. By default creates an in-org copy with an optional new name and server group assignments. Pass target_zone_ids to instead clone a patch policy into one or more zones/orgs in a single server-side call (patch policies only; mutually exclusive with name/server_groups).

Inputs: policy_id (required), name, request_id, server_groups, +1 more. Returns: object

create_data_extractCreate Data ExtractB
Destructive

Request a new data extract for bulk reporting. Returns the extract ID and initial status.

Inputs: extract_data (required), request_id. Returns: object

create_global_api_keyCreate Global API KeyA
Destructive

Create a global (account-scoped) API key. Returns metadata only — the key secret is never surfaced and cannot be retrieved via MCP.

Inputs: name (required), expires_at, request_id. Returns: object

create_policy_windowCreate Policy WindowA
Destructive

Create a new maintenance/exclusion window that prevents policy execution on the specified groups during the defined periods. All fields are required. The 'rrule' field does NOT accept generic RFC 5545; the upstream validator enforces exactly two grammars (live-verified 2026-06-05). For recurrence=once: FREQ=DAILY;UNTIL=YYYYMMDDTHHMMSSZ (compact UNTIL, no COUNT). For recurrence=recurring: FREQ=YEARLY;BYMONTH=<1-12>;BYDAY=<+N weekday> only — no other FREQ values, no WEEKLY. status accepts active|inactive; recurrence accepts once|recurring (lowercase on input).

Inputs: rrule (required), status (required), dtstart (required), recurrence (required), +8 more. Returns: object

create_server_groupCreate Server GroupC
Destructive

Create a new Automox server group. refresh_interval is the agent check-in/scan cadence in minutes (live-verified; spec range 240-1440).

Inputs: name (required), refresh_interval (required), parent_server_group_id (required), notes, +3 more. Returns: object

create_user_api_keyCreate User API KeyA
Destructive

Create an API key for a user. Returns metadata only — the key secret is never surfaced and cannot be retrieved via MCP.

Inputs: name (required), user_id (required), expires_at, request_id. Returns: object

create_webhookCreate WebhookA
Destructive

Create a new Automox webhook subscription. The response includes a signing secret that is ONLY shown once — save it immediately. Max 5 webhooks per organization. URL must be HTTPS.

Inputs: url (required), name (required), event_types (required), org_uuid, +1 more. Returns: object

create_zoneCreate ZoneB
Destructive

Create a new zone (organization) in the account. The zone access_key is never surfaced.

Inputs: name (required), request_id. Returns: object

decide_patch_approvalDecide Patch ApprovalB
Destructive

Approve or reject an Automox patch approval request.

Inputs: decision (required), approval_id (required), notes, request_id. Returns: object

delete_action_setDelete Action SetA
DestructiveIdempotent

Delete a single vulnerability remediation action set by ID. Action sets are console metadata (not endpoint state) and are reconstructable by re-uploading the source CSV.

Inputs: action_set_id (required), request_id. Returns: object

delete_action_sets_bulkDelete Action Sets BulkA
DestructiveIdempotent

Delete multiple vulnerability remediation action sets by ID (up to 100) in one atomic call. Action sets are reconstructable by re-uploading the source CSV.

Inputs: action_set_ids (required), request_id. Returns: object

delete_global_api_keyDelete Global API KeyA
DestructiveIdempotent

Permanently delete a global (account-scoped) API key by ID.

Inputs: key_id (required), request_id. Returns: object

delete_policyDelete PolicyB
DestructiveIdempotent

Permanently delete an Automox policy by ID.

Inputs: policy_id (required), request_id. Returns: object

delete_policy_windowDelete Policy WindowB
DestructiveIdempotent

Delete a maintenance/exclusion window permanently.

Inputs: window_uuid (required), org_uuid, request_id. Returns: object

delete_server_groupDelete Server GroupB
DestructiveIdempotent

Delete an Automox server group permanently.

Inputs: group_id (required), request_id. Returns: object

delete_user_api_keyDelete User API KeyA
DestructiveIdempotent

Permanently delete a user API key by user ID and key ID.

Inputs: key_id (required), user_id (required), request_id. Returns: object

delete_webhookDelete WebhookA
DestructiveIdempotent

Delete an Automox webhook subscription permanently.

Inputs: webhook_id (required), org_uuid, request_id. Returns: object

device_detailDevice DetailA
Read-onlyIdempotent

Return detailed information and recent activity for a device. Includes a compliance rollup (per-policy status counts: up_to_date / pending / needs_remediation, plus the policies needing remediation — only needs_remediation policies make a device non-compliant). uptime_minutes is sampled at the device's last full scan, so it can lag the current boot session.

Inputs: device_id (required), include_queue, include_packages, include_inventory, +1 more. Returns: object

device_health_metricsDevice Health MetricsA
Read-onlyIdempotent

Aggregate organization-wide device health statistics including managed/unmanaged breakdown, device status breakdown, compliance metrics, and check-in recency analysis. Compliance follows the platform rule: a device is non-compliant only when at least one policy needs remediation — pending work is tracked separately as devices_with_pending_policies. Use this for monitoring dashboards and getting a fleet-wide health overview.

Inputs: limit, group_id, include_unmanaged, max_stale_devices. Returns: object

device_search_typeaheadDevice Search TypeaheadA
Read-onlyIdempotent

Get typeahead suggestions for device search fields. Useful for discovering valid values when building advanced device queries.

Inputs: field (required), prefix (required), output_format.

discover_capabilitiesDiscover CapabilitiesA
Read-onlyIdempotent

Canonical inventory of Automox MCP tools, organized by domain — prefer this over cached manifests. With a domain: that domain's tools with name, description, and per-session availability (unavailable tools carry available=false; opt-in gated tools name their gated_by env var). With no domain: the domain index plus self-check totals (unique_tool_count, registered_tool_count, per-domain counts) and alias/cross-listing metadata; pass list_all_tools=true for a flat deduplicated tool list. Valid domains: devices, device_search, policies, policy_history, patches, groups, events, reports, audit, webhooks, worklets, data_extracts, vuln_sync, account, compound, policy_windows, splashtop. This directory excludes discover_capabilities itself.

Inputs: domain, list_all_tools. Returns: object

execute_policy_nowExecute Policy NowB
Destructive

Execute an Automox policy immediately for remediation (all devices or specific device).

Inputs: action (required), policy_id (required), device_id, request_id. Returns: object

get_accountGet AccountB
Read-onlyIdempotent

Get Automox account detail (id, name, type, timestamps).

Inputs: output_format.

get_account_userGet Account UserA
Read-onlyIdempotent

Get an account-scoped user record by UUID: status, account RBAC role, verification, and 2FA type. two_factor_authentication carries the literal string 'disabled' when 2FA is OFF (do NOT read it as a configured type); a null/absent value is ambiguous (may mean disabled or not-reported) — see metadata.field_notes.

Inputs: user_id (required), output_format.

get_action_set_detailGet Action Set DetailC
Read-onlyIdempotent

Get details for a specific vulnerability remediation action set. The status field is a processing-lifecycle string; observed live (2026-06-06) as building -> ready, with 'ready' the confirmed terminal value. 'active' is the spec example only and was not reproduced. See metadata.field_notes.

Inputs: action_set_id (required), output_format.

get_action_set_issuesGet Action Set IssuesA
Read-onlyIdempotent

Get vulnerability issues (CVEs) associated with an action set. Shows which vulnerabilities are being tracked for remediation.

Inputs: action_set_id (required), output_format.

get_action_set_solutionsGet Action Set SolutionsB
Read-onlyIdempotent

Get solutions for a vulnerability action set. Shows recommended patches or configurations to resolve vulnerabilities. Per-vulnerability severity and per-device status are coded strings with no API-enumerated value set (severity scale/ceiling spec-derived and unverified live; device status observed live transitioning not-started -> in_progress, note the mixed hyphen/underscore separators) — see metadata.field_notes.

Inputs: action_set_id (required), output_format. Returns: Structured output of the ``get_action_set_solutions`` tool.

get_cached_search_resultsGet Cached Search ResultsA
Read-onlyIdempotent

Retrieve cached server-side results for a previously-executed device search, keyed by search execution ID. Distinct from get_saved_search_results which re-executes a saved-search definition.

Inputs: search_id (required), page, limit, output_format.

get_compliance_snapshotGet Compliance SnapshotA
Read-onlyIdempotent

Combined view of non-compliant devices, fleet health metrics, and policy statistics. Answers 'What is our compliance posture?' in a single call.

Inputs: group_id, detail_limit. Returns: Structured output of ``get_compliance_snapshot``.

get_data_extractGet Data ExtractB
Read-onlyIdempotent

Get details for a data extract job. Returns status (per spec enum: queued | running | complete | failed | canceled | expired) and the is_completed boolean (use this for readiness). download_expires_at is the link-expiry timestamp; has_download_url=true means the CSV is currently downloadable. The presigned download URL itself is not returned to the model — retrieve it from the Automox console.

Inputs: extract_id (required), output_format.

get_device_assignmentsGet Device AssignmentsA
Read-onlyIdempotent

Get device-to-policy and device-to-group assignments.

Inputs: output_format.

get_device_by_uuidGet Device by UUIDA
Read-onlyIdempotent

Get the near-raw device payload by UUID (canonical /servers endpoint). Integer policy codes carry a status_label sibling (up_to_date / pending / needs_remediation — only needs_remediation makes a device non-compliant; see the compliance rollup). uptime_minutes is sampled at the device's last full scan, so it can lag the current boot session.

Inputs: device_uuid (required), output_format.

get_device_full_profileGet Device Full ProfileA
Read-onlyIdempotent

Complete device profile combining device detail, inventory summary, packages, and policy assignments in a single call. Inventory is summarized with key values per category. Packages capped at max_packages (default 25). pending_commands entries carry command_type and scheduled_time (the queued exec_time). Use get_device_inventory or list_device_packages for full data.

Inputs: device_id (required), detail_limit, max_packages. Returns: Structured output of ``get_device_full_profile``.

get_device_inventoryGet Device InventoryA
Read-onlyIdempotent

Retrieve detailed device inventory data including hardware, network, security, services, system, and user information. Optionally filter by category. Uses the Console API device-details endpoint.

Inputs: device_id (required), category. Returns: object

get_device_metadata_fieldsGet Device Metadata FieldsC
Read-onlyIdempotent

Get available fields for device queries. Returns the field names and types supported by the advanced device search API.

Inputs: output_format.

get_device_scheduled_windowsGet Device Scheduled WindowsB
Read-onlyIdempotent

Get upcoming scheduled maintenance periods for a specific device. Returns start/end times and window types. Optionally provide a future date limit (ISO 8601 UTC). start/end are derived occurrence times (the window stores dtstart+duration_minutes+rrule, not start/end); their timezone basis follows the window's use_local_tz.

Inputs: device_uuid (required), date, org_uuid. Returns: object

get_group_scheduled_windowsGet Group Scheduled WindowsA
Read-onlyIdempotent

Get upcoming scheduled maintenance periods for a server group. Returns start/end times and window types. Optionally provide a future date limit (ISO 8601 UTC). start/end are derived occurrence times (the window stores dtstart+duration_minutes+rrule, not start/end); their timezone basis follows the window's use_local_tz.

Inputs: group_uuid (required), date, org_uuid. Returns: object

get_patch_tuesday_readinessGet Patch Tuesday ReadinessA
Read-onlyIdempotent

Combined view of pre-patch report, pending approvals, and patch policy schedules. Answers 'Are we ready for Patch Tuesday?' in a single call. Policy schedule_days bitmasks carry a schedule_days_decoded sibling; schedule_time is a bare HH:MM string with no timezone marker. Per-device prepatch_report.devices[].highest_severity uses the vocabulary critical/high/medium/low/none/no_known_cves/unknown (full enum per spec; live-observed on this tenant: no_known_cves, high, critical, unknown). The projection emits the string 'unknown' for absent/undetermined severity — it never emits JSON null. 'no_known_cves' (patches carry no CVE) and 'unknown' (severity absent/undetermined, NOT inherently high risk) are distinct states and are never collapsed.

Inputs: group_id, detail_limit. Returns: Structured output of ``get_patch_tuesday_readiness``.

get_policy_windowGet Policy WindowA
Read-onlyIdempotent

Retrieve details for a specific maintenance/exclusion window by UUID. Returns status (lowercase active | inactive) and recurrence (UPPERCASE ONCE | RECURRING on read).

Inputs: window_uuid (required), org_uuid. Returns: object

get_saved_search_resultsGet Saved Search ResultsA
Read-onlyIdempotent

Execute a saved device search and retrieve its current device result set. Supports pagination via page + limit.

Inputs: saved_search_id (required), page, limit, output_format.

get_searchable_fieldsGet Searchable FieldsA
Read-onlyIdempotent

List searchable device fields grouped by scope, with per-field type metadata. Richer than get_device_metadata_fields (a flat field-name array) — use this to construct typed advanced-search queries.

Inputs: output_format.

get_search_scopesGet Search ScopesA
Read-onlyIdempotent

List available device-search scope options. Org-independent metadata describing the scopes (e.g., device, group, org) supported by the Advanced Device Search API.

Inputs: output_format.

get_server_groupGet Server GroupA
Read-onlyIdempotent

Get detailed information about a specific Automox server group. refresh_interval is the agent check-in/scan cadence in minutes (live-verified; spec range 240-1440) — see metadata.field_notes.

Inputs: group_id (required). Returns: object

get_upload_formatsGet Upload FormatsA
Read-onlyIdempotent

Get supported CSV upload formats for vulnerability remediation action sets.

Inputs: output_format.

get_userGet UserA
Read-onlyIdempotent

Get a single user by numeric ID, including org/server-group membership and RBAC roles. Secrets are never surfaced.

Inputs: user_id (required), output_format.

get_user_api_keyGet User API KeyA
Read-onlyIdempotent

Get one user API key by user ID and key ID. Returns metadata only — the secret is never exposed.

Inputs: key_id (required), user_id (required), output_format.

get_webhookGet WebhookB
Read-onlyIdempotent

Retrieve details for a specific Automox webhook subscription.

Inputs: webhook_id (required), org_uuid. Returns: object

get_worklet_detailGet Worklet DetailA
Read-onlyIdempotent

Get detailed information for a specific community worklet, including evaluation code, remediation code, and requirements.

Inputs: item_id (required), output_format.

get_zoneGet ZoneA
Read-onlyIdempotent

Get a single zone (organization) by UUID. The zone access_key is never surfaced.

Inputs: zone_id (required), output_format.

invite_user_to_accountInvite User to AccountC
Destructive

Invite a user to the Automox account with optional zone assignments.

Inputs: email (required), account_rbac_role (required), request_id, zone_assignments. Returns: object

list_account_rbac_rolesList Account RBAC RolesB
Read-onlyIdempotent

List the RBAC roles available in the Automox account.

Inputs: output_format.

list_data_extractsList Data ExtractsB
Read-onlyIdempotent

List data extract jobs for the Automox organization. Returns each job's id, type (e.g. patch-history | api-activity), status, the is_completed readiness boolean, and whether a download link is available (has_download_url).

Inputs: output_format.

list_device_packagesList Device PackagesA
Read-onlyIdempotent

List software packages on a specific Automox device. Returns package name, version, installed (boolean install state), repo, is_managed, and severity. There is no patch-status field in the response; install state comes from the installed boolean. severity is one of critical/high/medium/low/no_known_cves or JSON null (live-verified); none/unknown exist per spec but were not observed live. JSON null means no severity assessment was recorded (not a safety claim); no_known_cves means scanned with no known CVEs. See metadata.field_notes.severity. By default returns the complete package set (auto-paginated), so it is reliable for 'is package X installed?' checks. Pass an explicit page to fetch a single page instead.

Inputs: device_id (required), page, limit, output_format.

list_devices_for_policiesList Devices for PoliciesA
Read-onlyIdempotent

List the devices currently targeted by one or more policies (by policy UUID) — blast-radius assessment before executing or changing a policy. Read-only.

Inputs: policies (required), output_format.

list_eventsList EventsA
Read-onlyIdempotent

List Automox organization events with optional filters by policy, device, user, event name, or date range. For policy/patch events, data.status is the raw process exit code (arrives as string or int): '0' = success; negative values on Windows are NTSTATUS codes as signed 32-bit ints (e.g. '-1073741502' = 0xC0000142).

Inputs: page, limit, user_id, end_date, +6 more.

list_global_api_keysList Global API KeysA
Read-onlyIdempotent

List global (account-scoped) API keys. Returns key metadata (name, enabled, expiry) only — secrets are never exposed.

Inputs: output_format.

list_organizationsList OrganizationsA
Read-onlyIdempotent

List organizations visible to the API key, with device count, device limit, parent org, and trial end time. Useful for MSP/multi-org navigation, capacity posture, and trial warnings. A tier slug may be present per spec but has no defined feature ordering and is absent on some tenants — do not infer a paid-plan or capability ranking from it.

Inputs: page, limit, output_format.

list_org_api_keysList Org API KeysA
Read-onlyIdempotent

List API keys for the Automox organization. Returns key metadata (name, enabled, expiry) only — secrets are never exposed.

Inputs: output_format.

list_remediation_action_setsList Remediation Action SetsB
Read-onlyIdempotent

List vulnerability remediation action sets for the organization. Shows imported vulnerability data and remediation tracking.

Inputs: output_format.

list_saved_searchesList Saved SearchesB
Read-onlyIdempotent

List saved device searches from the Advanced Device Search API. Returns saved search names, queries, and metadata.

Inputs: output_format.

list_server_groupsList Server GroupsB
Read-onlyIdempotent

List all Automox server groups with their device counts and assigned policies. refresh_interval is the agent check-in/scan cadence in minutes (live-verified; spec range 240-1440) — see metadata.field_notes.

Inputs: page, limit, output_format.

list_user_api_keysList User API KeysA
Read-onlyIdempotent

List a user's API keys by user ID. Returns key metadata (name, enabled, expiry) only — secrets are never exposed.

Inputs: user_id (required), page, limit, output_format.

list_usersList UsersA
Read-onlyIdempotent

List users in the organization with name, email, and RBAC roles. Secrets (e.g. intercom_hmac) are never surfaced.

Inputs: page, limit, output_format. Returns: Structured output of the ``list_users`` tool.

list_webhook_deliveriesList Webhook DeliveriesA
Read-onlyIdempotent

List recent delivery attempts (status, latency, error) for a webhook -- delivery troubleshooting. Results are newest-first and cursor-paginated; optional startDate/endDate filters.

Inputs: webhook_id (required), limit, cursor, end_date, +3 more.

list_webhook_event_typesList Webhook Event TypesA
Read-onlyIdempotent

List all available Automox webhook event types with descriptions. Use this to see which events can trigger webhook deliveries.

Inputs: no named parameters. Returns: object

list_webhooksList WebhooksB
Read-onlyIdempotent

List all webhook subscriptions for the Automox organization. Supports cursor-based pagination.

Inputs: limit, cursor, org_uuid, output_format.

list_zonesList ZonesB
Read-onlyIdempotent

List the zones (organizations) in the Automox account.

Inputs: page, limit, output_format.

list_zones_for_userList Zones for UserB
Read-onlyIdempotent

List the zones (organizations) a given user belongs to.

Inputs: user_id (required), output_format.

list_zone_usersList Zone UsersC
Read-onlyIdempotent

List the users assigned to a given zone (by zone UUID).

Inputs: zone_id (required), page, limit, output_format.

noncompliant_reportNoncompliant ReportA
Read-onlyIdempotent

Retrieve the Automox non-compliant devices report showing devices that need attention due to policy failures or missing patches. Each failing policy includes 'reason_for_fail' (upstream failure text, may be truncated), 'severity', and 'type' so you can state why a device is non-compliant and prioritize across devices.

Inputs: limit, offset, group_id, output_format. Returns: Structured output of the ``noncompliant_report`` tool.

patch_approvals_summaryPatch Approvals SummaryA
Read-onlyIdempotent

Summarize patch approvals awaiting decision. Each approval covers one software package (title, software.version/os_family, CVE ids under cves) for a specific policy; manual_approval is true = approved, false = rejected, null = awaiting decision. The upstream record has no severity field, so severity_breakdown buckets approvals without one as 'unspecified' — use the CVE ids to judge urgency.

Inputs: limit, status, output_format. Returns: Structured output of the ``patch_approvals_summary`` tool.

policy_catalogPolicy CatalogA
Read-onlyIdempotent

List Automox policies with type and status summaries. schedule_days is a bitmask (decoded alongside as schedule_days_decoded); schedule_time is a bare HH:MM string with no timezone marker — device-local vs. org time depends on policy configuration, so the timezone cannot be inferred from this field alone.

Inputs: page, limit, include_stats, output_format, +1 more.

policy_compliance_statsPolicy Compliance StatsA
Read-onlyIdempotent

Retrieve per-policy compliance statistics for the organization. compliance_rate_percent is computed over evaluated devices (compliant + noncompliant) and is null when none have been evaluated; pending devices do not count against compliance and are reported separately via pending_devices / pending_rate_percent over all targeted devices.

Inputs: no named parameters. Returns: object

policy_detailPolicy DetailB
Read-onlyIdempotent

Retrieve configuration and recent history for a policy.

Inputs: policy_id (required), include_recent_runs. Returns: object

policy_execution_countsPolicy Execution CountsA
Read-onlyIdempotent

List fleet-wide policy execution counts over a time window: one row per policy with its run count, in a single round-trip. Answers 'which policies ran most last quarter?' without per-policy calls or client-side aggregation. Distinct from policy_run_count (single aggregate) and policy_runs_for_policy (per-run records for one policy).

Inputs: end_time, start_time, output_format.

policy_execution_timelinePolicy Execution TimelineB
Read-onlyIdempotent

Review recent executions for a policy. Each execution's device_outcomes are device counts per outcome (not run statuses); a run with no successes or failures but pending/not_included/remediation_not_applicable completed as a benign no-op or is still in progress, not a failure.

Inputs: policy_uuid (required), limit, org_uuid, report_days. Returns: object

policy_health_overviewPolicy Health OverviewB
Read-onlyIdempotent

Summarize recent Automox policy activity. status_breakdown counts RUNS by overall result; the 'no_success_or_failure' bucket means every device was pending/not_included/remediation_not_applicable (a benign no-op or still in progress), not an error.

Inputs: max_runs, org_uuid, window_days, top_failures. Returns: object

policy_history_detailPolicy History DetailB
Read-onlyIdempotent

Get policy history details by UUID, including run history and status. Each run's device_outcomes are device counts per outcome, not run statuses.

Inputs: policy_uuid (required), output_format, recent_runs_limit.

policy_run_countPolicy Run CountA
Read-onlyIdempotent

Get aggregate policy execution counts. Optionally filter by number of days to look back.

Inputs: days, output_format.

policy_run_detail_v2Policy Run Detail v2A
Read-onlyIdempotent

Get detailed per-device results for a specific policy run. Uses UUID-based queries and supports device name filtering. exit_code is the raw process exit code from the policy script (0 = success; negative values on Windows are NTSTATUS codes as signed 32-bit ints); result_status is the lowercase per-device outcome (e.g. success, failed).

Inputs: exec_token (required), policy_uuid (required), page, sort, +4 more.

policy_run_resultsPolicy Run ResultsB
Read-onlyIdempotent

Retrieve per-device results and output for a specific policy execution token. exit_code is the raw process exit code from the policy script (0 = success; negative values on Windows are NTSTATUS codes as signed 32-bit ints, e.g. -1073741502 = 0xC0000142); result_status is the lowercase per-device outcome string.

Inputs: exec_token (required), policy_uuid (required), page, sort, +5 more. Returns: object

policy_runs_by_policyPolicy Runs by PolicyA
Read-onlyIdempotent

Get policy runs grouped by policy for cross-policy comparison. Shows which policies have been running and their aggregate results.

Inputs: output_format.

policy_runs_for_policyPolicy Runs for PolicyA
Read-onlyIdempotent

Get execution runs for a specific policy by UUID. Optionally filter by number of days and sort order. Each run's device_outcomes are device counts per outcome, not run statuses.

Inputs: policy_uuid (required), sort, report_days, summary_only, +1 more.

policy_runs_v2Policy Runs v2A
Read-onlyIdempotent

List policy runs with time-range filtering, policy name/type filters, and result status filtering. Uses the Policy History v2 API for richer data than the standard policy execution timeline. Each run's device_outcomes (pending/success/failed/not_included/remediation_not_applicable/blocked) are DEVICE COUNTS per outcome for that run, not run statuses. result_status filters with any-device-with-this-outcome semantics: a run matches when AT LEAST ONE device had that outcome — it does NOT mean every device did (live-verified 2026-06-05: result_status='failed' returns runs with 1 failed device alongside 200+ not-failed). The same run can match multiple result_status values.

Inputs: page, sort, limit, end_time, +6 more.

prepatch_reportPrepatch ReportA
Read-onlyIdempotent

Retrieve the Automox pre-patch readiness report showing devices with pending patches before the next scheduled patch window. Per-device 'highest_severity' distinguishes 'no_known_cves' (patches carry no associated CVE — benign) from 'unknown' (severity undetermined). 'compliant' follows the platform rule: a device is non-compliant only when a policy needs remediation; pending work alone does not count against it.

Inputs: limit, offset, group_id, output_format. Returns: Structured output of the ``prepatch_report`` tool.

preview_policy_device_filtersPreview Policy Device FiltersA
Read-onlyIdempotent

Dry-run: preview which devices a policy's targeting would resolve to, before creating or updating the policy. Read-only — nothing is created or changed. server_groups is REQUIRED when device_filters is provided (the upstream preview cannot evaluate a filter-only target): pass the groups the policy will be assigned to, and the device_filters clauses ({field, op, value}) are applied within that scope.

Inputs: page, limit, output_format, server_groups, +1 more.

refresh_saved_search_cacheRefresh Saved Search CacheA
Idempotent

Force a re-cache of a saved device search's results when they may be stale. Triggers server-side recomputation; returns once queued.

Inputs: search_id (required), request_id. Returns: object

remove_user_from_accountRemove User from AccountB
DestructiveIdempotent

Remove a user from the Automox account by UUID.

Inputs: user_id (required), request_id. Returns: object

rotate_webhook_secretRotate Webhook SecretA
Destructive

Rotate the signing secret for an Automox webhook. The old secret is immediately invalidated. Save the new secret — it is only shown once.

Inputs: webhook_id (required), org_uuid, request_id. Returns: object

search_devicesSearch DevicesA
Read-onlyIdempotent

Search Automox devices by hostname (including custom name), IP, tag, severity of missing patches, or patch status (only 'missing' is supported). The severity filter is passed upstream to scope the result set (the requested values are echoed in metadata.filters.severity); it narrows the candidate set but does NOT guarantee every returned device currently has an outstanding patch at that severity (live, ~41% of severity=critical results report zero outstanding patches — likely already remediated or reassessed since the index was built). Per-device pending_patches is the device's total outstanding-patch count and is NOT scoped to the severity filter, so do not read it as the count matching the requested severity.

Inputs: tag, limit, managed, group_id, +5 more.

search_org_packagesSearch Org PackagesA
Read-onlyIdempotent

Search software packages across the Automox organization. Filter by managed status (include_unmanaged) or install state (awaiting: true = available but not installed, false = already installed -- per spec; note this is inverted from a naive reading of the word). awaiting is a request filter only, NOT a field in the response. Returns package name, version, is_managed, and severity. severity vocabulary is the same as list_device_packages: critical/high/medium/low/no_known_cves or JSON null observed live, with none/unknown present in the spec enum but unobserved. See metadata.field_notes.severity. The endpoint returns a bare list with no total, so returned_package_count is the count of packages on THIS page only (NOT a fleet-wide org total); when metadata.pagination.has_more is true, increment page to enumerate the rest.

Inputs: page, limit, awaiting, output_format, +1 more.

search_policy_windowsSearch Policy WindowsA
Read-onlyIdempotent

Search and list maintenance/exclusion windows for the Automox organization. Supports filtering by group UUIDs, status (active/inactive), and recurrence type (recurring/once). The recurrences filter is case-insensitive — tokens are coerced to the spec's UPPERCASE enum (ONCE/RECURRING) before sending, since the upstream expects uppercase (per spec, not live-verified). Supports pagination via page/size.

Inputs: page, size, sort, org_uuid, +5 more.

search_worklet_catalogSearch Worklet CatalogA
Read-onlyIdempotent

Search the Automox community worklet catalog. Returns worklet names, descriptions, categories, OS compatibility, and trust/availability signals (verified boolean, access tier such as premium, license_required). Catalog items have no status field. Use to discover pre-built evaluation and remediation scripts.

Inputs: query, output_format.

splashtop_device_statusSplashtop Device StatusA
Read-onlyIdempotent

List Splashtop installation and registration status for a device. Returns installation_status, registration_status, install_time, and any install error. Per the API spec (not live-verified by this server): installation_status=true means the client install completed (install_time is set); registration_status=true means the device has registered with Splashtop (a device identifier exists). These are independent — a device can be installed but not yet registered, so check both before assuming remote control is usable.

Inputs: device_uuid (required), output_format.

splashtop_force_disconnectSplashtop Force DisconnectA
DestructiveIdempotent

Force-disconnect ALL active Splashtop sessions on a device. Interrupts any in-progress operator work. Use sparingly.

Inputs: os_family (required), device_uuid (required), request_id. Returns: object

splashtop_get_attended_accessSplashtop Get Attended AccessA
Read-onlyIdempotent

Get the current attended-access requirement for a device. When requiredAttendedAccess is true, the end user must approve before a remote-control session can start.

Inputs: device_uuid (required), output_format.

splashtop_initiate_connectionSplashtop Initiate ConnectionA
Destructive

Generate a Splashtop deeplink (splashtop-sos://...) for an operator to start a remote-control session. The API does NOT start the session itself — the operator must open the returned URL in their local Splashtop RMM App, and end-user consent still applies if attended access is enabled on the device.

Inputs: os_family (required), device_uuid (required), connection_type (required), request_id, +1 more. Returns: object

splashtop_installSplashtop InstallA
Destructive

Install the Splashtop RMM client on a device. The install runs asynchronously. Use 'request_permission' to control whether the user is prompted at install time (distinct from per-session attended access). Requires a Remote Control subscription.

Inputs: os_family (required), device_uuid (required), request_id, account_type, +2 more. Returns: object

splashtop_session_statusSplashtop Session StatusA
Read-onlyIdempotent

List active Splashtop session capacity for a device. Returns can_start_new_session, current_sessions, and max_sessions. can_start_new_session reflects capacity only — per the spec examples it is true exactly when current_sessions < max_sessions. It is independent of attended-access consent: when attended access is enabled (see splashtop_get_attended_access) the end user must still approve before a session starts, so capacity alone is not sufficient to conclude a session can begin.

Inputs: device_uuid (required), account_type, output_format.

splashtop_set_attended_accessSplashtop Set Attended AccessA
DestructiveIdempotent

Enable or disable the end-user consent requirement for Splashtop sessions on a device. Setting this to false allows unattended sessions — review your organization's policy before doing so.

Inputs: device_uuid (required), required_attended_access (required), request_id. Returns: object

splashtop_set_bulk_attended_accessSplashtop Set Bulk Attended AccessA
DestructiveIdempotent

Bulk-set the attended-access requirement across many devices. Setting required_attended_access=false on a list disables end-user consent on all of them at once — confirm before use.

Inputs: device_uuids (required), required_attended_access (required), request_id. Returns: object

splashtop_uninstallSplashtop UninstallB
DestructiveIdempotent

Uninstall the Splashtop RMM client from a device and delete its registration + attended-access setting. Permanent.

Inputs: os_family (required), device_uuid (required), request_id. Returns: object

test_webhookTest WebhookA
Destructive

Send a test delivery to an Automox webhook endpoint. Returns success status, HTTP status code, and response time.

Inputs: webhook_id (required), org_uuid, request_id. Returns: object

update_deviceUpdate DeviceA
Idempotent

Update a single device's mutable attributes: custom_name, server_group_id, exception (policy-enforcement exclusion), tags, and ip_addrs. Fills the single-device gap that batch_update_devices (tags-only, bulk) does not cover — e.g. renaming a device or moving it to a server group. Supply only the fields you want to change; at least one is required.

Inputs: device_id (required), tags, ip_addrs, exception, +3 more. Returns: object

update_global_api_keyUpdate Global API KeyA
DestructiveIdempotent

Enable or disable a global (account-scoped) API key by ID.

Inputs: key_id (required), is_enabled (required), request_id. Returns: object

update_policy_windowUpdate Policy WindowA
DestructiveIdempotent

Update an existing maintenance window. Only dtstart is required; all other fields are optional for partial updates. status is lowercase active | inactive; recurrence is lowercase once | recurring on input. The 'rrule' field does NOT accept generic RFC 5545; the upstream validator enforces exactly two grammars (live-verified 2026-06-05). For recurrence=once: FREQ=DAILY;UNTIL=YYYYMMDDTHHMMSSZ (compact UNTIL, no COUNT). For recurrence=recurring: FREQ=YEARLY;BYMONTH=<1-12>;BYDAY=<+N weekday> only — no other FREQ values, no WEEKLY.

Inputs: dtstart (required), window_uuid (required), rrule, status, +9 more. Returns: object

update_server_groupUpdate Server GroupB
DestructiveIdempotent

Update an existing Automox server group. refresh_interval is the agent check-in/scan cadence in minutes (live-verified; spec range 240-1440).

Inputs: name (required), group_id (required), refresh_interval (required), parent_server_group_id (required), +4 more. Returns: object

update_userUpdate UserA
DestructiveIdempotent

Update a user's profile fields (firstname, lastname, email, tfa_type) by user ID. Passwords cannot be set through this tool.

Inputs: user_id (required), email, lastname, tfa_type, +2 more. Returns: object

update_user_api_keyUpdate User API KeyB
DestructiveIdempotent

Enable or disable a user API key by user ID and key ID.

Inputs: key_id (required), user_id (required), is_enabled (required), request_id. Returns: object

update_webhookUpdate WebhookA
DestructiveIdempotent

Update an existing Automox webhook. Only provided fields are changed (partial update). Can update name, URL, enabled status, or event types.

Inputs: webhook_id (required), url, name, enabled, +3 more. Returns: object

upload_action_setUpload Action SetA
Destructive

Upload a CSV-based vulnerability remediation action set. Provide the CSV as text in csv_content; source selects the format (generic | qualys | tenable | crowd-strike | rapid7). Call get_upload_formats first to see the required columns per source. The filename becomes the action set's display name. Returns the created action set (status is usually 'building' — processing is async).

Inputs: csv_content (required), source, filename, request_id. Returns: object

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3/5.0

    Scored across 122 tools

    Disambiguation2/5

    There are multiple clusters of highly similar tools, especially around saved searches (get_saved_search_results, run_saved_search, get_cached_search_results), policy runs (policy_runs_v2, policy_runs_for_policy, policy_execution_counts, policy_history_detail), and device detail (device_detail, get_device_by_uuid, get_device_full_profile). Even with detailed descriptions, an agent can easily select the wrong tool when several appear to execute or summarize the same underlying resource.

    Naming Consistency3/5

    Most resource CRUD families follow a consistent verb_noun pattern like list_server_groups, create_server_group, update_server_group, delete_server_group. However, the set mixes in noun-first names such as policy_detail, policy_catalog, prepatch_report, device_health_metrics, and patch_approvals_summary, plus awkward constructions like splashtop_get_attended_access and device_search_typeahead, making the overall convention inconsistent but still generally readable.

    Tool Count1/5

    122 tools is an extreme count for a single MCP server and far exceeds the point where the tool set becomes easy for an agent to navigate. Even though Automox is a broad platform, the server would be far more usable split into focused servers or consolidated into fewer, more composable tools.

    Completeness4/5

    The server provides impressively deep lifecycle coverage across devices, policies, device search, saved searches, server groups, policy windows, webhooks, users, API keys, patch approvals, data extracts, vulnerability action sets, and Splashtop remote control. Minor gaps exist, such as no update/delete for zones and no cancel for data extracts, but these are workable edge cases rather than core dead ends.

    Maintenance

    ActivitySlowing
    ResponsivenessUsually handled

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables AI assistants to manage Okta users, groups, applications, policies, and logs through natural language commands.
      Apache 2.0
    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables AI assistants to interact with NinjaOne through a decision-tree interface for managing devices, organizations, alerts, and tickets via natural language.
      Apache 2.0
    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables AI assistants to directly query and manage Autotask PSA data through natural language, including ticket triage, time entry creation, company lookups, project management, billing review, and more.
      Apache 2.0