Driving Nimbus with an agent

Driving Nimbus with an agent

Nimbus speaks the Model Context Protocol. Point any MCP client at it with a scoped token and it can define content types, write content, and manage the site — bounded by the token's scopes.

• Over HTTP at POST /api/v1/mcp, or locally over stdio with nimbus mcp.
• The server ships its own operating guide: initialize returns a brief, and resources/read serves nimbus://guide/core. Installed plugins ship their own guides too, so an agent knows how to drive each capability it touches.
• Every tool re-checks the token's scopes, and granting is subset-only — an agent can only ever do what its token already allows.

Connecting

Mint a scoped token, then talk to the endpoint over JSON-RPC 2.0 with a bearer header (quote the scopes so your shell doesn't expand the *):

nimbus token:create --name=agent --scopes='schema:write,*:read,*:write'

curl -X POST https://your-site/api/v1/mcp \
  -H "Authorization: Bearer nbt_your_token" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Call a tool the same way. For example, publish an entry once a posts collection exists:

{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{
  "name":"create_posts",
  "arguments":{"title":"Hello","status":"published","fields":{"body":"..."}}
}}

initialize returns the operating brief, and resources/read with nimbus://guide/core returns the full guide — so a connecting agent learns the workflow from the server itself. Both HTTP and the stdio transport (nimbus mcp) speak the same protocol.

Give an agent the narrowest token for the task, and it can build you a whole site.