CISO Assistant MCP Server – GRC Risk Automation
CISO Assistant includes a local Model Context Protocol server in its CLI folder for connecting MCP-compatible AI clients to a CISO Assistant instance. It lets agents query and, when permitted by the configured account, manage governance, risk, compliance, audit, asset, evidence, and third-party risk data through the CISO Assistant REST API.
Overview
CISO Assistant is an open-source governance, risk, and compliance platform from Intuitem. Its repository includes a CLI-based MCP server that runs locally and connects an AI assistant to a running CISO Assistant instance. The server is intended for Claude Desktop and other MCP-compatible clients, using the same CISO Assistant API permissions as the Personal Access Token configured for the integration.
What the MCP server enables
The MCP server exposes CISO Assistant workflows as structured tools. The official setup guide describes broad coverage across risk management, compliance audits, asset management, third-party risk management, EBIOS RM methodology, and related GRC data. The repository README also documents example MCP tools for retrieving risk scenarios, applied controls, and audit progress, while the setup guide lists examples such as showing risk assessments, listing imported frameworks, creating folders, adding risk scenarios, creating compliance assessments, reviewing gap analysis, and managing vendors.
When to use it
Use this MCP server when an AI agent needs live context from a CISO Assistant deployment rather than static security guidance. Practical workflows include summarizing high-risk scenarios, exploring assets in a folder, checking compliance audit progress, reviewing controls that are not yet implemented, preparing an assessment conversation, or helping a security team navigate GRC records through natural-language prompts.
Connection and authentication
The documented transport is stdio. The MCP client launches the local server with uv from the cloned ciso-assistant-community/cli directory, using uv --directory /path/to/ciso-assistant-community/cli run ca_mcp.py. Authentication is configured with a CISO Assistant Personal Access Token. The documented environment variables are TOKEN, API_URL, and VERIFY_CERTIFICATE; they can be placed in .mcp.env or supplied by the MCP client configuration.
Key considerations
The MCP server runs locally and calls the configured CISO Assistant API endpoint. Python 3.12+, uv, a running CISO Assistant instance, and a valid Personal Access Token are prerequisites. Production configurations should keep VERIFY_CERTIFICATE=true when using a valid certificate, avoid committing .mcp.env or .clica.env, and scope the token to the minimum permissions needed. Because tools can read or write GRC records depending on permissions, review client approval prompts and account access before enabling it for sensitive environments.
Supported Transports
stdio
Command: uv
Args:
--directory${CISO_ASSISTANT_CLI_DIR}runca_mcp.py
Frequently Asked Questions
- When should an AI agent use the CISO Assistant MCP server?
- Use it when an agent needs to work with live CISO Assistant GRC data, such as risk assessments, compliance audits, assets, controls, evidence, third-party risk records, or EBIOS RM studies, instead of relying on generic security knowledge.
- What does this MCP server add to an AI agent's capabilities?
- It gives the agent structured access to CISO Assistant tools for querying and managing GRC data through the CISO Assistant REST API, including risk, audit, control, asset, and vendor-related workflows supported by the local MCP server.
- What can the AI agent access or manage through this server?
- The official setup guide describes coverage for risk management, compliance audits, asset management, third-party risk management, EBIOS RM, and related records. Actual access depends on the configured CISO Assistant instance, the MCP tool set, and the permissions of the Personal Access Token used by the server.
- How is authentication configured for this MCP server?
- Generate a Personal Access Token in CISO Assistant, then configure the MCP server with `TOKEN`, `API_URL`, and `VERIFY_CERTIFICATE` either in the local `.mcp.env` file or through the MCP client's environment block. Never commit those environment files or raw token values to source control.
- Which transport should be used for this MCP server?
- Use the documented `stdio` transport. The CISO Assistant guide explains that the local stdio server avoids opening network ports, runs as a subprocess launched by the AI client, and sends outbound API calls from the user's machine to the configured CISO Assistant API URL.