threatlocker-mcp-server
An MCP server that exposes the ThreatLocker Portal API to AI assistants for querying and managing endpoint protection (computers, applications, policies, audit logs, approvals) over stdio or HTTP/SSE.
Computers — search/list, get details, check-in history, install info; isolate, lockdown, re-secure, baseline rescan, restart agent, rename/regroup, move to another org, delete records, restart org fleet, remove duplicates
Computer groups — list with computers/policies, dropdowns, per-OS filtering, lookup by install key, approval-workflow groups
Applications — search by name/hash/cert/country, get, ThreatLocker research, list file rules, match files, create/update apps, add/remove file rules, delete (soft or force)
Policies — get, list all or by application, create/update (full replace), delete, copy between groups, deploy pending changes, ringfencing/schedules/user-scoping options
Audit & investigations — unified action log search/detail/file history/policy conditions/testing details, portal system audit, health center, reports, saved searches (list/insert/delete)
Approvals — list/count/get pending requests, permit (two-step), reject, ignore, take ownership, storage/USB approvals, permit options and testing environment details
Maintenance mode — history, enable any mode (monitor/installation/learning/elevation/isolation/lockdown/etc.), end early, extend window
Organizations — list child orgs/tree, get auth key, orgs for moving computers, timezones, create child org, rotate auth key
Fleet management — scheduled agent version updates (list/search/schedule/abort), available agent versions, currently online devices
Network & storage — network access policy list/get/create, storage policy list/get, tags get/dropdown/update
Forensics — request and retrieve endpoint file upload requests
Safety/ops — read-only mode (
THREATLOCKER_READ_ONLY), retries/timeouts, pagination withfetchAllPages, JSON or markdown output, health/tools/REST endpoints in HTTP mode
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@threatlocker-mcp-server list all computers in my organization"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
ThreatLocker MCP Server
An MCP (Model Context Protocol) server for interacting with the ThreatLocker Portal API through Claude Desktop, Claude Code, or any MCP-compatible client.
About
This server exposes ThreatLocker Portal functionality as MCP tools, enabling AI assistants to query computers, applications, policies, audit logs, and more. It supports both local (stdio) and remote (HTTP/SSE) transports.
Current Status: Full read/write support for applications and policies. Set THREATLOCKER_READ_ONLY=true to enforce read-only mode.
Related MCP server: DelineaMCP
Disclaimer
USE AT YOUR OWN RISK
This software is provided "as is" without warranty of any kind. This is an unofficial, community-developed integration and is not affiliated with, endorsed by, or supported by ThreatLocker.
API keys are currently stored in plain text (in environment variables,
.envfiles, or MCP client config files). A more secure credential storage solution is planned for a future release.Always test in a non-production environment first
Review the source code before deploying
Monitor API usage and audit logs
The authors are not responsible for any damages, security incidents, or unintended actions resulting from use of this software
By using this software, you accept full responsibility for its use in your environment.
Protecting API Keys with ThreatLocker Storage Control
Since API keys are stored in plain text, you can use ThreatLocker's own Storage Control to restrict which applications can read the config files. This ensures that even if an unauthorized process runs on your machine, it cannot access the keys.
Files to protect:
File | Used By |
| MCP server (stdio mode) |
| Claude Desktop |
| Claude Code |
Recommended Storage Control policy:
In ThreatLocker Portal, navigate to Application Control > Storage Control
Create a Deny policy that blocks all applications from reading the config files listed above
Create Permit policies that allow only the specific applications that need access. Example:
node.exe/node— for the MCP server processClaude Desktop.exe/Claude Desktop— for Claude Desktopclaude— for Claude Code CLI
Apply the policies to the relevant computer group
This way, ThreatLocker prevents any other process from reading your API keys, even though they are stored in plain text.
Installation
Prerequisites
Node.js 26+ or Docker
ThreatLocker API key (generate in Portal)
Option 1: Docker (Recommended)
docker pull ghcr.io/bigfootbytes/threatlocker-mcp-server:latestOption 2: From Source
git clone https://github.com/BigfootBytes/threatlocker-mcp-server.git
cd threatlocker-mcp-server
npm install
npm run buildConfiguration
Claude Desktop / Claude Code
Add to your MCP config file:
Client | OS | Config Path |
Claude Desktop | macOS |
|
Claude Desktop | Windows |
|
Claude Desktop | Linux |
|
Claude Code | All | Project |
Docker configuration:
{
"mcpServers": {
"threatlocker": {
"command": "docker",
"args": ["run", "-i", "--rm", "ghcr.io/bigfootbytes/threatlocker-mcp-server:latest"],
"env": {
"THREATLOCKER_API_KEY": "your-api-key",
"THREATLOCKER_BASE_URL": "https://portalapi.g.threatlocker.com/portalapi",
"THREATLOCKER_ORG_ID": "optional-managed-org-id"
}
}
}
}Node.js configuration:
{
"mcpServers": {
"threatlocker": {
"command": "node",
"args": ["/path/to/threatlocker-mcp-server/dist/index.js"],
"env": {
"THREATLOCKER_API_KEY": "your-api-key",
"THREATLOCKER_BASE_URL": "https://portalapi.g.threatlocker.com/portalapi"
}
}
}
}Environment Variables
Variable | Required | Default | Description |
| Yes* | - | API key (stdio mode) |
| Yes* | - | Portal API URL |
| No | - | Managed organization ID |
| No |
| Transport mode: |
| No |
| HTTP server port |
| No |
| Logging: |
| No | - | CORS origins (comma-separated) |
| No | - | Set to |
| No |
| Retries for transient API failures. Writes retry only on 408/417/429, never after a 5xx, network error, or timeout |
| No |
| Per-attempt API request timeout in milliseconds |
*Required for stdio mode. HTTP mode uses per-request headers.
ThreatLocker API URLs
Environment | Base URL |
Production |
|
Beta |
|
Available Tools
CRUD Capabilities
Tool | Create | Read | Update | Delete | Description |
| - | :white_check_mark: | - | - | Query computers, check-ins, install info |
| - | :white_check_mark: | - | - | List groups, dropdowns |
| - | :white_check_mark: | - | - | Search apps, research details, files |
| - | :white_check_mark: | - | - | View policies by ID or application |
| - | :white_check_mark: | - | - | Unified audit logs, file history |
| - | :white_check_mark: | - | - | Pending approvals, permit details |
| - | :white_check_mark: | - | - | Child orgs, auth keys |
| - | :white_check_mark: | - | - | List and run reports |
| - | :white_check_mark: | - | - | Computer maintenance history |
| - | :white_check_mark: | - | - | Scheduled agent updates |
| - | :white_check_mark: | - | - | Portal audit logs, health center |
| - | :white_check_mark: | - | - | Network and policy tags |
| - | :white_check_mark: | - | - | Storage control policies |
| - | :white_check_mark: | - | - | Network access control policies |
| - | :white_check_mark: | - | - | Available ThreatLocker agent versions |
| - | :white_check_mark: | - | - | Currently online/connected devices |
Tool Details
Tool | Actions |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
HTTP Mode (Remote Server)
For remote deployments, run in HTTP mode:
docker run -d -p 8080:8080 -e TRANSPORT=http ghcr.io/bigfootbytes/threatlocker-mcp-server:latestEndpoints
Method | Endpoint | Auth | Description |
GET |
| No | Health check |
GET |
| No | List available tools |
GET |
| Yes | SSE stream (Claude Desktop) |
POST |
| Session | SSE client messages |
POST |
| Yes | Streamable HTTP MCP |
POST |
| Yes | Direct REST API |
Authentication Headers
Header | Required | Description |
| Yes | ThreatLocker API key |
| Yes | Portal API base URL |
| No | Managed organization ID |
Claude Remote Configuration
Streamable HTTP via mcp-remote (Claude Desktop):
Claude Desktop does not yet support Streamable HTTP natively. Use mcp-remote as a proxy:
{
"mcpServers": {
"threatlocker": {
"command": "npx",
"args": [
"mcp-remote",
"https://your-server.example.com/mcp",
"--header",
"Authorization:${THREATLOCKER_API_KEY}",
"--header",
"X-ThreatLocker-Base-URL:${THREATLOCKER_BASE_URL}"
],
"env": {
"THREATLOCKER_API_KEY": "your-api-key",
"THREATLOCKER_BASE_URL": "https://portalapi.g.threatlocker.com/portalapi"
}
}
}
}SSE (legacy):
{
"mcpServers": {
"threatlocker": {
"url": "https://your-server.example.com/sse",
"headers": {
"Authorization": "your-api-key",
"X-ThreatLocker-Base-URL": "https://portalapi.g.threatlocker.com/portalapi"
}
}
}
}Development
npm install # Install dependencies
npm run build # Compile TypeScript
npm test # Run tests
npm run dev # Watch modeLicense
GPL-3.0 - see LICENSE for details.
Available Tools
18 toolsaction_logThreatLocker Action LogARead-onlyIdempotent
Query ThreatLocker unified audit logs.
The action log records all application control events: permits, denies, network access, file operations, PowerShell execution, elevation requests, and more. This is your primary tool for investigating what happened on endpoints.
Common workflows:
Find all denies in last 24 hours: action=search, startDate="...", endDate="...", actionId=99
Find denies on a specific computer: action=search, ..., hostname="COMPUTER-NAME"
Find network blocks: action=search, ..., actionType=network, actionId=2
Find PowerShell executions: action=search, ..., actionType=powershell
Get details of a specific event: action=get, actionLogId="..."
Track a file's history across all computers: action=file_history, fullPath="C:\path\to\file.exe"
Aggregate by user to find who's triggering denies: action=search, ..., groupBys=[1]
Get file download details: action=get_file_download, actionLogId="..."
Get policy conditions for permit: action=get_policy_conditions, actionLogId="..."
Get testing environment details: action=get_testing_details, actionLogId="..."
Pitfalls:
onlyTrueDenies/simulateDeny only filter when used alone or together; they force actionId=99 internally. "True" deny = enforced block; "simulated" = would-have-blocked on a Monitor/Learning computer.
When calling get/get_file_download, pass the sourceTableId matching the row the eActionLogId came from (default 2=DenyActionLog will miss permit/baseline/eventlog events).
groupBys takes at most 2 fields; prefer it over fetching raw rows for aggregation.
username/deviceType are NOT supported search filters here (the V2 endpoint ignores them); pivot on hostname/fullPath/policyId or use groupBys=[1] to break down by user.
Permissions: View Unified Audit. Pagination: search action is paginated (use fetchAllPages=true to auto-fetch all pages). Performance: always use date filters — queries without startDate/endDate can be very slow on large organizations. Use groupBys to aggregate instead of fetching all raw rows. Key response fields: actionLogId, fullPath, processPath, hostname, username, actionType, policyName, applicationName.
Related tools: computers (find computer IDs), applications (identify apps), approval_requests (handle denied software)
Inputs: action (required), endDate, actionId, fullPath, +18 more. Returns: data, error, success, pagination
applicationsThreatLocker ApplicationsADestructive
Search, inspect, create, update, and delete ThreatLocker applications.
Applications are collections of file rules (hashes, paths, certificates) that define what software is allowed or denied. ThreatLocker comes with built-in applications for common software, and you can create custom ones.
Common workflows:
Find an application by name: action=search, searchText="Chrome"
Find apps by file hash: action=search, searchBy=hash, searchText="abc123..."
Find apps by certificate: action=search, searchBy=cert, searchText="Microsoft"
Get ThreatLocker research on an app: action=research, applicationId="..."
List files in an application: action=files, applicationId="..."
Find apps actively permitted: action=search, permittedApplications=true
Find recently created custom apps: action=search, category=1, orderBy=date-created
Find matching apps by file properties: action=match, hash="...", path="...", cert="..."
Get apps for maintenance mode: action=get_for_maintenance
Get app for network policy: action=get_for_network_policy, applicationId="..."
Create custom application: action=create, name="My App", osType=1
Update application metadata: action=update, applicationId="...", name="...", osType=1
Manage a child org's app: add managedOrganizationId="child-org-guid" to create/update
Add file rules to application: action=add_file, applicationId="...", osType=1, fileRules=[{hash:"..."}, {fullPath:"...", cert:"..."}]
Remove file rules from application: action=remove_file, applicationId="...", applicationFileIds=[7111524894, 7111524907] (get IDs via action=files)
Delete application (no policies): action=delete, applications=[{applicationId:"...", name:"...", organizationId:"...", osType:1}]
Force delete (with policies): action=delete_confirm, applications=[...]
Pitfalls:
Hash-only file rules must contain only the hash (no path/cert); file paths need double-escaped backslashes in JSON.
create makes metadata only — add file rules in a follow-up add_file call; then build a policy and deploy it.
remove_file needs applicationFileId values from action=files first.
Built-in applications take policy precedence over custom apps.
Permissions: Edit Application Control Applications. Pagination: search and files actions are paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: applicationId, name, osType, computerCount, policyCount. Research fields: concernRating, reviewRating, categories, countriesWhereCodeCompiled.
Related tools: policies (see policies using this app), action_log (see app activity), approval_requests (pending approvals for this app)
Inputs: action (required), cert, hash, name, +30 more. Returns: data, error, success, pagination
approval_requestsThreatLocker Approval RequestsADestructive
Query ThreatLocker approval requests.
When users encounter blocked software and request access, it creates an approval request. Admins review these requests to decide whether to permit the software by creating policies.
Common workflows:
List pending requests: action=list, statusId=1
Get pending request count: action=count
Find requests for a specific user: action=list, searchText="username"
Get request details: action=get, approvalRequestId="..."
Get file info for download/analysis: action=get_file_download_details, approvalRequestId="..."
Get permit options (apps, groups): action=get_permit_application, approvalRequestId="..."
Get storage request details: action=get_storage_approval, approvalRequestId="..."
Approve a request: action=permit. Two-step — call get_permit_application first, round-trip its opaque "json" blob into permitJson, then pick permitMode + policyLevel. Payload-verified, NOT live-tested: validate in a non-prod org before relying on it.
Request statuses: 1=Pending (needs review), 4=Approved, 6=Not Learned (learning mode), 10=Ignored, 12=Added to Application, 13=Escalated (from Cyber Heroes), 16=Self-Approved
Pitfalls:
list defaults to newest-first (isAscending=false) — the right default for triaging the pending queue.
Permitting a request is a two-step flow: call get_permit_application first and round-trip its opaque "json" blob; don't synthesize it.
Before approving a Built-In matching app, confirm the file isn't a shared DLL matching unrelated apps (you'd permit the whole built-in).
Permissions: View Approvals, Approve for Entire Organization/Group/Single Computer. Pagination: list action is paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: approvalRequestId, username, fullPath, actionType, statusId, computerName, requestDateTime.
Related tools: action_log (see the deny event), applications (find matching apps), policies (create permits)
Inputs: action (required), osType, ruleId, orderBy, +51 more. Returns: data, error, success, pagination
computer_groupsThreatLocker Computer GroupsARead-onlyIdempotent
List and inspect ThreatLocker computer groups.
Computer groups organize computers and define policy scope. Policies are applied to groups, not individual computers. The "global" group (includeGlobal=true) permits applications across all groups.
Common workflows:
Get all groups with computers: action=list, includeAllComputers=true
Get group dropdown for UI/selection: action=dropdown
Get groups across organizations (MSP): action=dropdown_with_org, includeAvailableOrganizations=true
Filter by OS type: osType=1 (Windows), 2 (macOS), 3 (Linux)
Get groups for approval workflow: action=get_for_permit
Get group by install key: action=get_by_install_key, installKey="..."
Permissions: Super Admin (for list), Edit Computers, Edit Computer Groups, View Computers. Key response fields: computerGroupId, name, osType, computerCount, organizationId.
Related tools: computers (list computers in groups), policies (policies applied to groups)
Inputs: action (required), osType, installKey, hideGlobals, +14 more. Returns: data, error, success, pagination
computersThreatLocker ComputersADestructive
Query and inspect ThreatLocker computers.
Common workflows:
Find computers by logged-in user: action=list, searchBy=2, searchText="username"
Find computers by IP: action=list, searchBy=4, searchText="192.168.1.100"
List computers needing review: action=list, kindOfAction="NeedsReview"
Get computer details by ID: action=get, computerId="..."
View check-in history: action=checkins, computerId="..."
Get installation info for new deployments: action=get_install_info
Rename / re-group a computer: action=edit, computerId="...", computerGroupId="...", name="..."
Move a computer to another org: action=move_org, computerId="...", computerGroupId="...", organizationId="...", osType=1, targetComputerGroupId="...", targetOrganizationId="..."
Remove computers from the Portal: action=delete, deleteComputers=[{computerId, computerName, organizationId}] (same org; does NOT uninstall)
Restart every agent in the org: action=restart_org (includeChildOrganizations=true also hits child orgs)
Remove duplicate records: action=remove_duplicate
Pitfalls:
get returns the editable computer record, not live protection state; read current mode/isolation from list results or maintenance_mode history.
This is the triage entry point: find a box here, grab its computerId/organizationId/computerGroupId, then hand off to maintenance_mode, approval_requests, or action_log.
Permissions: View Computers, Edit Computers (for modifications), Install Computers (for install info). Pagination: list and checkins actions are paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: computerId, computerName, computerGroupName, lastCheckin, action (Secure/Installation/Learning/MonitorOnly), threatLockerVersion.
Related tools: computer_groups (manage groups), maintenance_mode (maintenance history), action_log (audit events)
Inputs: action (required), name, osType, endDate, +30 more. Returns: data, error, success, pagination
maintenance_modeThreatLocker Maintenance ModeADestructive
Query ThreatLocker maintenance mode history for computers.
Maintenance mode temporarily changes a computer's protection level. Types include:
Monitor Only (1): Logs but doesn't block (audit mode)
Installation Mode (2): Allows new software installs, auto-learns new applications
Learning Mode (3): Monitors and records software usage without blocking
Tamper Protection Disabled (6): Allows ThreatLocker service changes
Common workflows:
View maintenance history for a computer: action=get_history, computerId="..."
Audit who put computers in installation mode: check history across computers
Maintenance mode history shows who enabled it, when, duration, and what applications were learned during that time.
Pitfalls:
Learning Mode (3) requires a "Default - (Group Name)" Default Deny policy to exist in the group, or it silently does nothing.
Isolation (14) and Lockdown (15) require ThreatLocker Detect and Agent >= 8.2.
usersList entries are "DOMAIN\USERNAME" and only apply when allUsers=false; default window is 1 hour if no end time is given.
Permissions: Edit Computers, Manage Application Control Installation Mode, Manage Application Control Learning Mode. Pagination: get_history is paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: maintenanceModeId, maintenanceTypeId, displayName, startDateTime, endDateTime, addedBy, endedBy.
Related tools: computers (get computer IDs, see current mode), computer_groups (group-level modes)
Inputs: action (required), computerId (required), notes, allUsers, +12 more. Returns: data, error, success, pagination
network_access_policiesThreatLocker Network Access PoliciesA
Query ThreatLocker network access control policies.
Network access policies define firewall rules for endpoints — controlling which applications can make or receive network connections, and to which destinations (IPs, ports, domains).
Common workflows:
List all network access policies: action=list
Search by name: action=list, searchText="RPC"
Filter by computer group: action=list, appliesToId="group-id"
Get policy details by ID: action=get, networkAccessPolicyId="..."
Pitfalls:
Network Control has no Global policy level (unlike application control).
Tag-based rules need the tag label + id resolved via the tags tool (dropdown); parent-org tags use the "ParentOrg\TagName" format.
direction: 1=Inbound, 2=Outbound; policyActionId: 1=Permit, 2=Deny.
Permissions: Edit Network Control Policies, View Network Control Policies. Pagination: list action is paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: networkAccessPolicyId, name, computerGroupName, isEnabled, applicationName.
Related tools: policies (application control policies), computer_groups (where policy applies), tags (network tags used in policies)
Inputs: action (required), name, ports, status, +16 more. Returns: data, error, success, pagination
online_devicesThreatLocker Online DevicesARead-onlyIdempotent
Query ThreatLocker online devices.
Returns devices currently connected and reporting to the ThreatLocker platform. Useful for real-time visibility into which endpoints are active.
Common workflows:
Check how many devices are online right now: action=list
Verify a specific computer is connected: action=list, then search results for hostname
Monitor fleet connectivity after a network change: action=list, compare count to computers tool total
Paginate through large device lists: action=list, pageNumber=2, pageSize=100
Permissions: View Computers. Pagination: list action is paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: computerName, computerGroupName, lastCheckin, ipAddress.
Related tools: computers (full inventory with details, modes, groups), computer_groups (group membership and structure)
Inputs: action (required), orderBy, pageSize, pageNumber, +3 more. Returns: data, error, success, pagination
organizationsThreatLocker OrganizationsADestructive
Query ThreatLocker organizations.
Organizations are the top-level containers in ThreatLocker. MSPs have a parent organization with child organizations for each client. Enterprises may have organizations per business unit or location.
Common workflows:
List child organizations: action=list_children
Search for a client org: action=list_children, searchText="client name"
List all nested children (full tree): action=list_children, includeAllChildren=true
Get installation auth key: action=get_auth_key
Get orgs available for moving computers: action=get_for_move_computers
Provision a client org: action=timezones (pick an id) → action=create_child, displayName="...", timezoneId="..." → get_auth_key → deploy
Rotate the org auth key: action=rotate_auth_key (DESTRUCTIVE — invalidates the old key and breaks existing deploy scripts)
The organizationId is needed for many API calls (policies, applications, etc.) to scope the request to a specific organization.
Pitfalls:
get_auth_key returns the install/auth key used to deploy agents and to resolve groups via computer_groups get_by_install_key.
Permissions: View Organizations, Edit Organizations, Super Admin - Child. Pagination: list_children is paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: organizationId, name, displayName, dateAdded, computerCount.
Related tools: computers (computers in org), computer_groups (groups in org), policies (policies in org)
Inputs: action (required), name, domains, options, +17 more. Returns: data, error, success, pagination
policiesThreatLocker PoliciesADestructive
Manage ThreatLocker policies.
Use list_all to search policies by computer group / org / filter without an applicationId; use list_by_application when you already have an applicationId.
Policies define what applications can run on which computer groups. A policy links an application (set of file rules) to a computer group with an action (permit/deny/ringfence).
Common workflows:
Get policy details by ID: action=get, policyId="..."
List all policies for an application: action=list_by_application, applicationId="...", organizationId="..."
Find policies for a specific group: action=list_by_application, applicationId="...", organizationId="...", appliesToId="group-id"
Include deny policies in results: action=list_by_application, ..., includeDenies=true
Create new policy: action=create, name="...", applicationIds=["..."], computerGroupId="...", osType=1, policyActionId=1
Update policy (full replace - get first!): action=update, policyId="...", name="...", applicationIds=["..."], computerGroupId="...", osType=1, policyActionId=1
Delete policies: action=delete, policyIds=["..."], organizationId="..."
Copy policies between groups: action=copy, osType=1, policyIds=["..."], sourceAppliesToId="...", sourceOrganizationId="...", targetAppliesToIds=["..."]
Deploy pending changes: action=deploy, organizationId="..."
IMPORTANT: After create/update/delete/copy, deploy changes with action=deploy to push to computers. IMPORTANT: Update is a full replace — use action=get first to read current values, then provide ALL fields.
Policy actions: Permit (allow), Deny (block), Ringfence (allow but restrict network/storage access)
Pitfalls:
Precedence is first-match-wins (Global > Global Group > Entire Org > Computer > Computer Group). New policies land at the bottom unless orderBefore=true.
monitorMode=1 (Secured) creates an explicit deny that overrides Learning Mode; monitorMode=2 is Monitor Only.
allowRequest/killRunningProcesses are only valid with policyActionId=2 (Deny).
Ringfence requires policyActionId=6 + ringfencingOptions (5 restrict flags + rf* arrays). rfFilePolicy permission: permit 1=read-only/2=read+write, deny 1=deny-write/2=deny-read+write. rfNetworkPolicy.server uses "tag:".
policySchedules requires policyScheduleStatus=2; networkExclusions pairs with restrictNetworkAccess=true.
update is FULL-REPLACE: get first, then resend ALL nested arrays (ringfencingOptions/policySchedules/networkExclusions/userGroups/parentProcessIdList/requestEmailAddressesList) you want to keep — omitting one removes it. Payload-verified, NOT live-tested.
Scoping: userGroups needs allUserGroups=false; deviceType needs allDevices=false; parentProcessIdList needs parentRestrictionEnabled=true (apps must match the policy osType); applicationSelection=1 (all apps) needs a name containing "Permit All" or starting "Default - "; notifyOnRequest/requestEmailAddressesList are Deny-only (policyActionId=2).
Permissions: View Application Control Policies, Edit Application Control Policies. Pagination: list_by_application is paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: policyId, name, policyActionId, applicationId, computerGroupId, isEnabled.
Related tools: applications (what the policy permits), computer_groups (where policy applies), action_log (see policy enforcement)
Inputs: action (required), name, filter, osType, +42 more. Returns: data, error, success, pagination
reportsThreatLocker ReportsARead-onlyIdempotent
Query and run ThreatLocker reports.
Access pre-built and custom reports configured in the ThreatLocker portal. Reports provide aggregated views of security data across your organization.
Common workflows:
List all available reports: action=list
Run a specific report: action=get_data, reportId="..." (get IDs from list action first)
Review security posture: list reports, then run relevant compliance or audit reports
Export data for external analysis: run a report and process the returned data
Permissions: View Reports. Key response fields: reportId, name, description, reportData (dynamic columns per report type).
Related tools: action_log (raw audit events), system_audit (portal audit trail), computers (device inventory)
Inputs: action (required), endDate, reportId, startDate, +4 more. Returns: data, error, success, pagination
saved_searchesThreatLocker Saved SearchesADestructive
Manage saved Unified Audit / investigation searches.
A saved search has two parts: searchData (the reproducible parameters, as JSON) and saveParameters (a human-readable label). saveSearchPageId selects which page the search belongs to (valid values 1-10; e.g. the Unified Audit page).
Common workflows:
List saved searches for a page: action=list, saveSearchPageId=
Save a search: action_log build_search_string (→ saveParameters label) → action=insert, saveSearchId=, saveSearchPageId=, organizationId="...", searchName="...", saveParameters="Any Deny", searchData={the search params object}
Delete a saved search: action=delete, saveSearchId="..."
Pitfalls:
searchData is REQUIRED and must be valid JSON (object or JSON string) — the API deserializes it and returns an opaque 500 if it is missing or malformed. (Verified live 2026-07-06.)
saveParameters is the display label from action_log build_search_string (e.g. "Any Deny").
saveSearchPageId must be a valid page type (1-10); 0/other values return "Page type wasn't found".
insert requires you to supply a fresh saveSearchId GUID.
Permissions: View Unified Audit.
Related tools: action_log (build_search_string produces saveParameters)
Inputs: action (required), datetime, username, searchData, +8 more. Returns: data, error, success, pagination
scheduled_actionsThreatLocker Scheduled ActionsA
Query ThreatLocker scheduled agent actions.
Scheduled actions are pending operations on ThreatLocker agents, primarily version updates. Updates are batched and scheduled within maintenance windows to avoid disruption.
Common workflows:
List all scheduled actions: action=list
Search with filters: action=search, organizationIds=["..."], computerGroupIds=["..."]
Get scheduled action details: action=get, scheduledActionId="..."
Get available targets for scheduling: action=get_applies_to
Scheduled action types: Version Update (scheduledType=1).
Pitfalls:
isAscending is inverted by the API: true (or omitted) sorts descending (high to low); set false for ascending.
search (GetByParameters) is keyed by scheduledId — pass it to filter to a specific scheduled action's computers; use list for the top-level set.
Permissions: Edit Computers, Edit Computer Groups, View Computers. Pagination: search action is paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: scheduledAgentActionId, scheduledType, scheduledDateTime, computerName, computerGroupName, status.
Related tools: computers (see current versions), computer_groups (target groups for updates), organizations (filter by org)
Inputs: action (required), osType, orderBy, abortAll, +18 more. Returns: data, error, success, pagination
storage_policiesThreatLocker Storage PoliciesARead-onlyIdempotent
Query ThreatLocker storage control policies.
Storage policies define rules for file and folder access on endpoints — controlling which applications can read, write, or execute from specific storage locations (local drives, USB devices, network shares).
Common workflows:
List all storage policies: action=list
Search by name: action=list, searchText="USB"
Filter by computer group: action=list, appliesToId="group-id"
Get policy details by ID: action=get, storagePolicyId="..."
Pitfalls:
Read-only tool: storage policy creation/editing is not available via the public API (no documented write endpoint).
Storage policies are first-match top-down — permits must be ordered above denies.
Permissions: View Storage Control Policies, Edit Storage Control Policies. Pagination: list action is paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: storagePolicyId, name, policyType, osType, computerGroupName, isEnabled.
Related tools: policies (application control policies), computer_groups (where policy applies), applications (what the policy permits)
Inputs: action (required), osType, pageSize, pageNumber, +6 more. Returns: data, error, success, pagination
system_auditThreatLocker System AuditARead-onlyIdempotent
Query ThreatLocker portal audit logs.
System audit tracks administrator actions in the ThreatLocker portal: logins, policy changes, approvals, configuration modifications. This is different from action_log which tracks endpoint events.
Common workflows:
Find all logins in date range: action=search, startDate="...", endDate="...", auditAction=Logon
Find failed login attempts: action=search, ..., auditAction=Logon, effectiveAction=Denied
Find changes by a specific admin: action=search, ..., username="admin@company.com"
Find policy modifications: action=search, ..., auditAction=Modify, details="policy"
Get health center dashboard: action=health_center, days=7
Search health center by location: action=health_center, searchText="lat:X&long:Y"
Audit actions: Create (new objects), Delete (removals), Logon (portal access), Modify (changes), Read (views). Supports * wildcard in text fields.
Permissions: View System Audit, View Health Center. Pagination: search and health_center actions are paginated (use fetchAllPages=true to auto-fetch all pages). Key response fields: systemAuditId, username, action, effectiveAction, details, ipAddress, dateTime.
Related tools: action_log (endpoint events, not portal events), organizations (filter by org)
Inputs: action (required), days, details, endDate, +12 more. Returns: data, error, success, pagination
tagsThreatLocker TagsA
Query ThreatLocker tags for network and policy management.
Tags are reusable labels for IP addresses, domains, ports, or other network identifiers. They simplify policy management by letting you reference "CRM Servers" instead of listing individual IPs.
Common workflows:
List all available tags: action=dropdown
Include ThreatLocker built-in tags: action=dropdown, includeBuiltIns=true
Get tag details by ID: action=get, tagId="..."
Update tag membership: action=update, tagId="...", organizationId="..." (full-object replace — get first, resend ALL item arrays you want to keep; there is no insert/delete endpoint)
Tags are used in:
Network Control policies (allow/deny traffic to tagged destinations)
Ringfencing (restrict app network access to tagged resources)
Storage Control (restrict file access to tagged paths)
Parent organization tags appear as "parentOrgName\tagName" format.
Pitfalls:
Use dropdown to get the label+value (tagId) needed when building network/ringfence policy payloads.
Parent-organization tags use the "ParentOrg\TagName" format.
Permissions: Edit Network Control Policies, Manage Tags, Edit Application Control Policies. Key response fields: tagId, name, tagType, values (IP/domain/port entries).
Related tools: policies (use tags in policy rules), applications (ringfence with tags)
Inputs: action (required), name, tagId, active, +12 more. Returns: data, error, success, pagination
upload_requestsThreatLocker Upload RequestsA
Request and retrieve forensic file uploads from endpoints.
Common workflows:
Request a file upload: action=insert, uploadRequestId=, organizationId="...", computerId="...", shA256="..." (or filepath)
Retrieve an upload request: action=get, uploadRequestId="..."
Pitfalls:
The SHA-256 field is literally "shA256" (odd casing) in the API.
insert requires you to supply a fresh uploadRequestId GUID.
Writes are payload-verified, NOT live-tested.
Permissions: View Unified Audit / forensics.
Related tools: action_log (locate the file event), computers (get computerId)
Inputs: action (required), hash, shA256, filename, +6 more. Returns: data, error, success, pagination
versionsThreatLocker VersionsARead-onlyIdempotent
Query available ThreatLocker agent versions.
Returns all agent versions available in the portal, including which are enabled for installation, which is the default for new groups, and when each was released.
Common workflows:
List all available versions: action=list
Find the latest version: action=list, look for highest version number with isEnabled=true
Check if a specific version is still available: action=list, search results for version string
Identify the default version for new computer groups: action=list, look for isDefault=true
Plan upgrade rollouts: action=list, compare to installed versions from computers tool
Permissions: Edit Computers, Edit Computer Groups, View Computers, Install Computers. No pagination — returns all versions in a single response. Key response fields: label (version string), value (version ID), isEnabled, dateTime (release date), isDefault, OSTypes.
Related tools: computers (see installed versions per machine), scheduled_actions (schedule version updates), computer_groups (group-level version settings)
Inputs: action (required), fetchAllPages, response_format. Returns: data, error, success, pagination
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 18 tools
Each tool maps to a distinct ThreatLocker domain—computers, groups, applications, policies, logs, approvals, organizations, reports, maintenance, scheduling, tags, storage/network policies, versions, online devices, saved searches, and upload requests. Potential overlaps like system_audit vs action_log and policies vs storage_policies/network_access_policies are clearly differentiated by both naming and description.
All tools follow the same snake_case resource-noun pattern: computers, computer_groups, applications, policies, action_log, maintenance_mode, storage_policies, etc. While actions are passed via an 'action' parameter rather than verb-prefixed tool names, the convention is uniform and easy to predict.
18 tools is above the ideal 3-15 range, but the count is justified by the breadth of the ThreatLocker platform—each tool covers a distinct functional area. It feels slightly heavy rather than bloated, and there are no redundant tools that could be merged.
Core lifecycle coverage is strong for applications and policies (create/update/delete/deploy), and there are solid querying tools for logs, approvals, and computers. However, several areas are read-only or lack management operations: computer_groups cannot be created/edited/deleted, storage_policies and network_access_policies are read-only, maintenance_mode is history-only, and scheduled_actions cannot be created or cancelled.
Maintenance
Related MCP Connectors
Official Microsoft MCP Server to query Microsoft Entra data using natural language
An MCP server that provides an API to LLMs to manage their JumpCloud resources.
MCP server for Pentest-Tools.com: run scans, manage findings and reports via your preffered LLM.
MCP server that lets AI assistants use all OneSchema features exposed via the public API.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceMCP server for ThreatLocker — zero-trust application allowlisting, approval requests, audit logs1Apache 2.0

DelineaMCPofficial
AlicenseNot gradedqualityBmaintenanceMCP server for the Delinea Secret Server and Platform APIs, enabling AI agents to manage secrets, users, groups, folders, roles, and access requests through natural language commands.50MIT- AlicenseNot gradedqualityCmaintenanceMCP server that enables AI-powered assessment of Active Directory on-premises environments by exposing AD data as queryable tools for LLMs like Claude.MIT
- AlicenseCqualityBmaintenanceAn MCP server that connects AI assistants to the ThreatLocker Portal API with 44 tools for managing computers, approvals, action logs, tags, maintenance mode, and more across single-org and multi-tenant setups.4465 PyPI2MIT