Agents / Claude Code

Claude Code

Recommended

Anthropic's coding agent, built for deep, autonomous multi-step work in your terminal.

The official plugin installs the MCP server and a set of slash commands together, in one step.

What you need

Claude Code installed, with plugin support
Node.js (for npx)
A syntica account (or create one during setup)

Install

Run once, from anywhere:

$ npx @syntica/mcp setup

This installs both the MCP server and the slash commands as a single Claude Code plugin, with no manual config file edits. Restart Claude Code afterward to activate it. The same command also finds and connects any other supported agent on your machine (Codex, Cursor, Windsurf, Gemini CLI, Copilot CLI), and Claude Code is the one we recommend: it is the only agent with a skill runtime, so the flows below arrive as real commands.

Available commands

/syntica:onboard

Authenticate, connect your git remote, scaffold docker-compose.yml if needed, configure secrets, and push your code.

/syntica:deploy

Push your latest code, trigger a deployment, and poll until it's live.

/syntica:debug

Inspect running services, tail logs, and restart a service. Needs the developer role or above.

/syntica:connections

See which systems this application has been granted — a mailbox, Stripe, Slack, SharePoint — so you write against the credentials it is actually given instead of inventing your own.

/syntica:manage

Check what you're being charged, resize an environment, and delete, restore or rename an application.

/syntica:security

View the validation prompts your agent runs against your code before you deploy — any role can do that. Adding, editing or removing your organisation's own prompts needs the developer role or above. Syntica does not run them for you.