Wazuh MCP Server
Wazuh MCP Server gives an agent access to the Wazuh SIEM so hosts, ports, packages and processes can be queried in plain language.
- 96 ★ on the repository
- Last commit
- AGPL-3.0
Extended access
Why this level
- Host details and Wazuh credentials reach the agent and the model.
- In HTTP mode without a token and network limits the server exposes infrastructure data.
Install
- Works in
Manual install
python -m venv .venv && source .venv/bin/activate
pip install git+https://github.com/socfortress/wazuh-mcp-server.gitThe first README option, installing from GitHub.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Host details and Wazuh credentials reach the agent and the model.
- In HTTP mode without a token and network limits the server exposes infrastructure data.
This is third-party code. Review the repository files before installing.
What it does
Wazuh MCP Server links a Wazuh manager to a language model. An agent gets the agent list with status filters and syscollector data for a given host: open ports, installed packages and running processes. The server refreshes JWT authorisation itself. It runs over stdio or Streamable HTTP, and network mode requires `MCP_AUTH_TOKEN`. The `WAZUH_READ_ONLY` mode and the `WAZUH_DISABLED_TOOLS` and `WAZUH_DISABLED_CATEGORIES` variables switch off unneeded tools.
Who it is for. SOC engineers and administrators who review host state in Wazuh and want to question the data in natural language.
Good fit when
- You need to see quickly which ports a suspicious host is listening on.
- You need to find hosts that run a specific package version.
- An engineer asks an agent to compare process lists across several machines.
- Access is granted through a dedicated Wazuh account with minimal rights.
Not a fit when
- The Wazuh manager is unreachable from the network where the agent runs.
- Infrastructure data cannot be passed to a language model.
- You need detection rules and incident response without a human.
Example request
List active Wazuh agents and for one of them show listening tcp ports and processes.MCP
- Transport
- stdio, http
- Authentication
- API key
Environment variables
WAZUH_PROD_URLrequired- Wazuh manager API address.
WAZUH_PROD_USERNAMErequired- Wazuh username.
WAZUH_PROD_PASSWORDrequired, secret- Wazuh user password.
MCP_AUTH_TOKENsecret- Token for HTTP mode.
Limitations
The toolset is narrow: the README lists agents, ports, packages and processes and has no ready alert workflows. Credentials sit in `.env`, so use a dedicated minimal-rights account with TLS verification on. Host data enters the model context. The README has inconsistencies: some tool descriptions mention capabilities missing from the current version, so compare the list with your client's `tools/list`. Set `WAZUH_PROD_SSL_VERIFY=true` to verify the manager certificate.
How to disable. Remove the MCP entry from the client configuration and stop the server. Revoke the Wazuh account created for it and delete the `.env` file.
FAQ
Can the server change Wazuh state?
In the current version every tool only reads data. `WAZUH_READ_ONLY=true` additionally leaves out tools that change state.
How do I run it over the network?
Run `MCP_AUTH_TOKEN=$(openssl rand -hex 32) wazuh-mcp-server --transport http --host 0.0.0.0 --port 8000`. The server answers at `/mcp`.
Is a dedicated account needed?
Yes. The README advises a Wazuh user with minimal permissions and TLS verification in production.
Which licence applies?
AGPL-3.0.
How do I disable unneeded tools?
List them in `WAZUH_DISABLED_TOOLS` or list categories in `WAZUH_DISABLED_CATEGORIES`.
Discussion
No questions or reviews yet.
Related
MCP serversMCP access to Wazuh for investigating alerts and vulnerabilities, checking SIEM health and invoking response tools
The official Zscaler MCP server: 402 tools for ZIA, ZPA, ZDX and other services, read-only by default
A CLI for auditing agent MCP servers: inventory, evidence per server and a CI policy where unmeasured never counts as clean
Security scanner for MCP servers that looks for hardcoded keys, tool poisoning and injection flaws in repository code.
The official MCP exposes Bitwarden vault operations and, with separate credentials, organization administration.
A bridge between an MCP client and a Kali Linux terminal so the agent can run nmap, sqlmap, hydra and other tools for authorized pentesting and CTF
README in short
The documentation covers installation from GitHub, environment-variable configuration, tool filtering and production security guidance.