Developer Documentation

Secureframe MCP Server

Connect an AI assistant to your Secureframe workspace and query or update your compliance data through the Model Context Protocol (MCP).

Overview

The Secureframe MCP server allows you to access and manage your Secureframe data — controls, tests, vendors, frameworks, risks, and more — as Model Context Protocol tools. Any MCP-capable client (Claude Code, Claude Desktop, Cursor, and others) can connect, letting you ask an assistant to look things up and take actions on your behalf.

Every request runs as a specific Secureframe user in a specific company and respects that user's existing permissions — the assistant can only see and do what you can.

Connect to your region

Use the URL that matches your Secureframe data region:

Region	MCP URL
United States	`https://mcp.secureframe.com/`
United Kingdom	`https://mcp-uk.secureframe.com/`

The examples below use the US URL — substitute your regional URL if applicable.

Authentication

There are two ways to authenticate. Both run every request as your Secureframe user, with your existing permissions.

Sign in with Secureframe recommended

Point your client at the MCP URL with no credentials at all. The first time it connects, it opens your browser, you sign in to Secureframe and approve access, and the client receives its own token. There is nothing to generate, copy, or store yourself, and no secret ends up in a config file.

This is standard OAuth 2.1, so most MCP clients — Claude Code, Claude Desktop, Cursor, MCP Inspector — handle it automatically: they discover the authorization server from the MCP URL and register themselves. See Connecting a client.

API key

For setups with no browser — CI jobs, scripts, servers — authenticate with a Secureframe REST API key and secret. Generate a key in Secureframe → Company Settings → API keys, then send the key and secret in an Authorization header, separated by a space:

Authorization: <API_KEY> <API_SECRET>

Treat your API secret like a password. Anyone with it can act as your user in Secureframe. Prefer supplying it through environment variables rather than pasting it into a shared config file — for example:

export SECUREFRAME_API_KEY="your-api-key"
export SECUREFRAME_API_SECRET="your-api-secret"

Then reference those variables when you connect, as shown below.

Connecting a client

Claude Code

Add the server with a single command — no credentials:

claude mcp add --transport http secureframe https://mcp.secureframe.com/

Then run /mcp inside Claude Code, select secureframe, and choose Authenticate. Your browser opens to Secureframe to sign in and approve access, and the connection completes on its own.

--transport http is required — without it, the CLI treats the argument as a local command to launch. --scope is optional: it controls where the config is saved — the default (local) applies to the current project only, while --scope user makes the server available across all your projects.

Other MCP clients

For clients configured with JSON (Claude Desktop, Cursor, etc.), point an HTTP MCP server at the endpoint:

{
  "mcpServers": {
    "secureframe": {
      "type": "http",
      "url": "https://mcp.secureframe.com/"
    }
  }
}

The client prompts you to sign in the first time it connects.

MCP servers are loaded when the client starts. After adding the server, restart your client so it picks up the new connection.

Connecting with an API key instead

If your client can't open a browser, supply the API key header directly. For Claude Code, using the environment variables from above:

claude mcp add --transport http secureframe https://mcp.secureframe.com/ \
  --header "Authorization: ${SECUREFRAME_API_KEY} ${SECUREFRAME_API_SECRET}"

Or, for a JSON-configured client:

{
  "mcpServers": {
    "secureframe": {
      "type": "http",
      "url": "https://mcp.secureframe.com/",
      "headers": {
        "Authorization": "<YOUR_API_KEY> <YOUR_API_SECRET>"
      }
    }
  }
}

A client configured with an Authorization header never starts the sign-in flow — it uses the key for every request.

What happens when you sign in

The first time a client connects without a key, it walks you through a short browser flow:

Sign in to Secureframe. Whatever you normally use works — email and password, SSO, or a social login, including any multi-factor step. If you are already logged into Secureframe, this is skipped.
Choose a company, if you belong to more than one that can connect MCP clients. With a single eligible company, it's selected for you and you never see this step.
Approve access. The consent screen names the client requesting access, the account you're signed in as, and the company the connection will act in. Approving grants the client the mcp scope — permission to use the Secureframe API as you.

Your browser then returns to the client, which stores its own token and is ready to use. You won't be asked again unless access is revoked.

The client is only ever granted your own access. It can see and change exactly what you can in the Secureframe app, in the company you selected — nothing more.

Who can connect a client

You can authorize a client for a company when your membership there is active and any one of the following is true:

you are an admin or super admin;
your role allows you to create API keys; or
you already own an active API key.

Sessions

An access token is valid for one hour. Clients renew it silently in the background, so you aren't asked to sign in every hour.
The renewal credential rotates each time it's used and stays valid until access is revoked.

Available tools

These tools wrap the Secureframe REST API. Any query filters and parameters available on an API endpoint are also available on its matching tool — see the full API documentation for details. Connected MCP clients receive each tool's parameter schema automatically, so your assistant knows what inputs a tool accepts without any extra setup.

112 tools across 41 categories. Expand a category to see its tools.

Cloud Resource3 tools

get_cloud_resource read

Get a Cloud Resource

list_cloud_resources read

List Cloud Resources

update_cloud_resource write

Update a Cloud Resource

Cloud Resource Framework Asset Scope2 tools

create_cloud_resource_framework_asset_scope write

Create Framework Asset Scope

list_cloud_resource_framework_asset_scopes read

List Framework Asset Scopes

Comment5 tools

create_comment write

Create a Comment

delete_comment write

Delete a Comment

get_comment read

Get a Comment

list_comments read

List Comments

update_comment write

Update a Comment

Control2 tools

get_control read

Get a Control

list_controls read

List Controls

Custom Integration1 tool

data_custom_integration write

Publish data

Device2 tools

get_device read

Get a Device

list_devices read

List Devices

Device Framework Asset Scope2 tools

create_device_framework_asset_scope write

Create Framework Asset Scope

list_device_framework_asset_scopes read

List Framework Asset Scopes

Evidence2 tools

get_evidence read

Get an Evidence

list_evidences read

List Evidence

File Upload1 tool

create_file_upload write

Stage a file so its bytes go straight to storage rather than through this API, and get back an id to attach it with. Uploading a file takes three steps, and only the first and third are tools — the middle one you make yourself. 1. Call this tool with the file's `filename`, `byte_size` and `checksum`. The response contains a `url`, a `headers` object, and an `id`. 2. PUT the file's bytes to that `url`, sending every entry in `headers` as a header, unaltered. The request body is the file's contents as they are on disk — raw bytes, not base64, not multipart, not wrapped in JSON — so there is nothing to encode or convert. Storage rejects the PUT unless the bytes match the `byte_size` and `checksum` declared in step 1, so declare them from the file you are actually sending. 3. Pass the `id` as `upload_id` to the tool that attaches it: `create_test_evidence`, `create_user_evidence` or `update_trust_center_request`. Each `id` is redeemable once; attaching the same file again means staging it again from step 1. A file must be 32 MB or smaller. A larger `byte_size` is refused here, in step 1, before you have spent anything on the upload itself. The two halves of the handshake expire apart, and the response dates both. The `url` stops being accepted 15 minutes after staging, at `url_expires_at`; the `id` stays redeemable for an hour, until `id_expires_at`. Bytes that have already landed can therefore still be attached after the URL is dead, but a batch of uploads staged up front must all be PUT inside that first 15 minutes.

Framework2 tools

get_framework read

Get a Framework

list_frameworks read

List Frameworks

Framework Requirement2 tools

get_framework_requirement read

Get a Framework Requirement

list_framework_requirements read

List Framework Requirements

Integration Connection3 tools

archive_integration_connection write

Archive an Integration Connection

get_integration_connection read

Get an Integration Connection

list_integration_connections read

List Integration Connections

Knowledge Base Answer4 tools

create_knowledge_base_answer write

Create a Knowledge Base Answer

delete_knowledge_base_answer write

Delete a Knowledge Base Answer

get_knowledge_base_answer read

Get a Knowledge Base Answer

update_knowledge_base_answer write

Update a Knowledge Base Answer

Knowledge Base Question4 tools

create_knowledge_base_question write

Create a Knowledge Base Question

delete_knowledge_base_question write

Delete a Knowledge Base Question

get_knowledge_base_question read

Get a Knowledge Base Question

update_knowledge_base_question write

Update a Knowledge Base Question

POA&M Item5 tools

create_poam_item write

Create a POA&M item

discard_poam_item write

Discard a POA&M item

get_poam_item read

Get a POA&M item

list_poam_items read

List POA&M items

update_poam_item write

Update a POA&M item

Policy2 tools

get_policy read

Get a Policy

list_policies read

List Policies

Repository3 tools

get_repository read

Get a Repository

list_repositories read

List Repositories

update_repository write

Update a Repository

Repository Framework Asset Scope2 tools

create_repository_framework_asset_scope write

Create Framework Asset Scope

list_repository_framework_asset_scopes read

List Framework Asset Scopes

Risk2 tools

get_risk read

Get a Risk

list_risks read

List Risks

SSP Duty5 tools

create_ssp_duty write

Create an SSP Duty

delete_ssp_duty write

Delete an SSP Duty

get_ssp_duty read

Get an SSP Duty

list_ssp_duties read

List SSP Duties

update_ssp_duty write

Update an SSP Duty

SSP Duty Role4 tools

create_ssp_duty_role write

Create an SSP Duty Role

delete_ssp_duty_role write

Delete an SSP Duty Role

get_ssp_duty_role read

Get an SSP Duty Role

list_ssp_duty_roles read

List SSP Duty Roles

SSP Policy5 tools

create_ssp_policy write

Create an SSP Policy

delete_ssp_policy write

Delete an SSP Policy

get_ssp_policy read

Get an SSP Policy

list_ssp_policies read

List SSP Policies

update_ssp_policy write

Update an SSP Policy

SSP Report3 tools

create_ssp_report write

Create an SSP Report

get_ssp_report read

Get an SSP Report

list_ssp_reports read

List SSP Reports

SSP Report Assessment Objective3 tools

get_ssp_report_assessment_objective read

Get an SSP Report Assessment Objective

list_ssp_report_assessment_objectives read

List SSP Report Assessment Objectives

update_ssp_report_assessment_objective write

Update an SSP Report Assessment Objective

SSP Report Section3 tools

get_ssp_report_section read

Get an SSP Report Section

list_ssp_report_sections read

List SSP Report Sections

update_ssp_report_section write

Update an SSP Report Section

SSP Report Section Block3 tools

get_ssp_report_section_block read

Get an SSP Report Section Block

list_ssp_report_section_blocks read

List SSP Report Section Blocks

update_ssp_report_section_block write

Update an SSP Report Section Block

SSP Role5 tools

create_ssp_role write

Create a SSP Role

delete_ssp_role write

Delete a SSP Role

get_ssp_role read

Get a SSP Role

list_ssp_roles read

List SSP Roles

update_ssp_role write

Update a SSP Role

SSP Vendor5 tools

create_ssp_vendor write

Create an SSP Vendor

delete_ssp_vendor write

Delete an SSP Vendor

get_ssp_vendor read

Get an SSP Vendor

list_ssp_vendors read

List SSP Vendors

update_ssp_vendor write

Update an SSP Vendor

Security Questionnaire1 tool

create_security_questionnaire write

Create a Security Questionnaire

Task2 tools

get_task read

Get a Task

list_tasks read

List Tasks

Test4 tools

create_test write

Create a Test

get_test read

Get a Test

list_tests read

List Tests

update_test write

Update a Test

Test Evidence1 tool

create_test_evidence write

Attach an evidence file to a Test. The bytes are not sent here. Stage the file first with the `create_file_upload` tool, which hands back a `url` and an `id`; PUT the file's raw bytes to that `url`; then call this tool with that `id` as `upload_id`. The bytes must already be in storage by the time you call this — an `upload_id` whose PUT never happened is refused rather than attached empty. `create_file_upload` documents the size limit and the two expiry windows. Each `upload_id` is redeemable once. Attaching the same file to a second Test means staging it again.

Test Export1 tool

create_test_export write

Create a Test Export

Test Export Reading1 tool

get_test_export_reading read

Get a Test Export

Third Party Risk Management Vendor3 tools

archive_third_party_risk_management_vendor write

Archive a Third Party Risk Management Vendor.

get_third_party_risk_management_vendor read

Get a Third Party Risk Management Vendor.

list_third_party_risk_management_vendors read

List Third Party Risk Management Vendors

Trust Center Request3 tools

get_trust_center_request read

Get a Trust Center Request

list_trust_center_requests read

List Trust Center Requests

update_trust_center_request write

Update a Trust Center Request: approve or reject the resources it asks for, set its document security, or send a custom response. Most calls attach no file at all. To attach a signed NDA agreement, the bytes are not sent here. Stage the file first with the `create_file_upload` tool, which hands back a `url` and an `id`; PUT the file's raw bytes to that `url`; then call this tool with that `id` as `upload_id`. The bytes must already be in storage by the time you call this — an `upload_id` whose PUT never happened is refused rather than attached empty. `create_file_upload` documents the size limit and the two expiry windows. Each `upload_id` is redeemable once. Attaching the same file to a second request means staging it again.

User3 tools

get_user read

Get a User

list_users read

List Users

update_user write

Update a User

User Account3 tools

get_user_account read

Get a User Account

link_user_account write

Link a User Account

list_user_accounts read

List User Accounts

User Evidence1 tool

create_user_evidence write

Attach an evidence file to a User. The bytes are not sent here. Stage the file first with the `create_file_upload` tool, which hands back a `url` and an `id`; PUT the file's raw bytes to that `url`; then call this tool with that `id` as `upload_id`. The bytes must already be in storage by the time you call this — an `upload_id` whose PUT never happened is refused rather than attached empty. `create_file_upload` documents the size limit and the two expiry windows. Each `upload_id` is redeemable once. Attaching the same file to a second User means staging it again.

User Security Settings1 tool

list_user_security_settings read

Get user security settings

Vendor3 tools

archive_vendor write

Archive a Vendor

get_vendor read

Get a Vendor

list_vendors read

List Vendors