Purple AI MCP Server
The official SentinelOne MCP server in read-only mode: Purple AI questions, PowerQuery, alerts, vulnerabilities and inventory.
- 103 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The SentinelOne token exposes alert, vulnerability and inventory data of your organization.
- Hashes, domains and IPs are sent to the external VirusTotal and cve-search.org services.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/purple-ai-mcp-serverDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The SentinelOne token exposes alert, vulnerability and inventory data of your organization.
- Hashes, domains and IPs are sent to the external VirusTotal and cve-search.org services.
This is third-party code. Review the repository files before installing.
What it does
The SentinelOne MCP server connects an MCP client to a SentinelOne console in read-only mode. The agent asks Purple AI questions, runs PowerQuery over events in the data lake, searches alerts, vulnerabilities and misconfigurations, and looks up inventory objects: endpoints, cloud resources, identities and network devices. Public CVE database search and file, URL, domain and IP analysis through VirusTotal with your own key are available as well. The server runs over `stdio`, `sse` and `streamable-http`.
Who it is for. SOC analysts and security engineers who have a SentinelOne console and a service token scoped to one account or site.
Good fit when
- Alert investigation: notes, history and related events in one conversation with the agent.
- Triage of fresh vulnerabilities by CVE with data from your own environment.
- An inventory question: which endpoints or cloud resources match a condition.
- Checking a file hash or domain in VirusTotal without leaving the chat.
Not a fit when
- You need to change data in the console: the server only reads and does not close alerts or isolate hosts.
- The token is global or spans several sites: only single-account or single-site tokens are supported.
- The server is meant to be exposed on a network without a proxy: it has no authentication of its own.
Example request
Show the latest alerts and, for the newest one, print its notes and change history.MCP
- Transport
- stdio, http, sse
- Authentication
- API key
Environment variables
PURPLEMCP_CONSOLE_TOKENrequired, secret- SentinelOne console service token for one account or site.
PURPLEMCP_CONSOLE_BASE_URLrequired- URL of your SentinelOne console.
PURPLEMCP_VT_API_KEYsecret- VirusTotal key for the threat_intel tools.
Limitations
A SentinelOne console and a service token from the Service Users section are required. The token is limited to one account or site, so several sites need several servers. The server has no authentication of its own: in network modes put it behind a reverse proxy or load balancer, while `/health`, `/ready` and `/ping` bypass the check. VirusTotal needs `PURPLEMCP_VT_API_KEY`, and `threat_intel_search` also needs a premium API. For production the authors advise pinning the install to a commit hash.
How to disable. Remove the `purple-mcp` entry from the client configuration, stop the server process or container and revoke the service token in the SentinelOne console.
FAQ
How do I connect Purple AI MCP to Claude Code?
Run `claude mcp add --transport stdio purple-mcp`, pass `PURPLEMCP_CONSOLE_TOKEN` and `PURPLEMCP_CONSOLE_BASE_URL` through `--env` and add `uvx --from git+https://github.com/Sentinel-One/purple-mcp.git purple-mcp --mode stdio` after `--`. Codex has the same form with `codex mcp add`.
Can the server change data in SentinelOne?
No. The authors describe it as read-only: the account and its objects cannot be changed through it.
Which token does Purple AI MCP need?
A service token with account-level or site-level permissions, not global. Create it in the console under Policy and Settings, User Management, Service Users.
How do I pin the server version?
Append `@<commit-hash>` to the repository URL in `uvx --from`. The authors advise it for production because of supply-chain risk.
Which transports does the server support?
`stdio`, `sse` and `streamable-http`. The mode is set with the `--mode` flag or the `PURPLEMCP_TRANSPORT_MODE` variable.
Discussion
No questions or reviews yet.
Related
MCP serversAn MCP server that gives the agent vulnerability triage: NVD data, EPSS, the CISA KEV catalog, MITRE ATT&CK mapping and one-call CVE risk scoring
Security Detections MCP: a detection rule database
Security Detections MCP
An MCP server that merges Sigma, Splunk, Elastic and KQL rules into one database for finding detections and ATT&CK gaps
Security scanner for MCP servers that looks for hardcoded keys, tool poisoning and injection flaws in repository code.
An MCP server for security document review: an agent submits files and the findings stay drafts until a person decides
MCP access to Wazuh for investigating alerts and vulnerabilities, checking SIEM health and invoking response tools
MCP-Shield finds your MCP configs, connects to servers and scans tool descriptions for hidden commands, exfiltration channels and tool shadowing
README in short
The vendor maintains the project. It runs via `uvx`, Docker, Amazon Bedrock AgentCore and ECS, and cloud guides cover load balancers. Connection examples cover Claude Desktop, Claude Code, Codex and Zed.