Connect an agent
Connect your own AI agent, such as Claude, or any MCP client to CISO360AI's Model Context Protocol (MCP) server, from its connector directory or with an API key, to work with your data under your role.
Connect from your AI tool's directory
- In your AI tool, open the connector directory and add CISO360AI.
- Sign in to CISO360AI when asked.
- Review what the connector asks for, then allow access.
To confirm, ask your agent "which projects can you see?"
Connect any MCP client with an API key
-
Create an API key of the Agent type with the fewest capabilities your agent needs.
-
In your MCP client, add an HTTP server and send the key as a bearer token:
{ "type": "http", "url": "https://api.ciso360.ai/mcp", "headers": { "Authorization": "Bearer <your-api-key>" } }
The key decides which organisation the agent can reach, so use one key for each organisation. How the server is shaped is in the API and MCP reference.
What an agent can do
| Connected agent | |
|---|---|
| Reads | Engagements, findings, vulnerabilities, assets, technologies, risks, frameworks, controls, and cloud configuration, in the projects your role reaches. Scan results are totals until you verify your domain |
| Acts in | What your role allows and, for an Agent key, the capabilities you chose. Deleting an asset or releasing a report is filed as a proposal under Tasks › Inbox; account and organisation tasks stay in the app |
| Held and removed under | Privacy Policy · Security · Stop an agent |
Did this answer your question?
Connect your AI agent
How your own AI agent connects to CISO360AI over MCP, the API keys it uses, and where the live API and MCP schemas are.
API keys
An API key lets an AI agent, script or other tool act in CISO360AI with its owner's role in one organisation; create, scope, rotate and revoke keys, and stop an agent.