> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.cessy.ai/agents-in-cessy/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.cessy.ai/_mcp/server.

# Agents In Cessy

> Use Design MCP for builder agents and Runtime MCP for published app agents without mixing design-time and runtime authority.

Cessy supports agents at two different layers.

| Layer       | MCP server  | Main job                                                                                                            |
| ----------- | ----------- | ------------------------------------------------------------------------------------------------------------------- |
| Design-time | Design MCP  | Help builders inspect, validate, change, preview, and publish app design.                                           |
| Runtime     | Runtime MCP | Help users operate a published app through generated commands, projections, channels, users, DMs, and page refresh. |

This separation is important. A runtime support agent should not need design-time mutation tools. A builder agent should not assume runtime state without reading the published contracts and environment.

## Design-time agent loop

A good builder agent starts with:

```text
get_app
get_domain_model
list_errors
get_reference
```

Then it proposes a workspace plan, makes scoped edits, validates, and summarizes runtime impact before promotion.

Use [Connect Design MCP](/start/connect-design-mcp) before asking Claude, Codex, ChatGPT, or another compatible MCP client to help with builder work.

## Runtime agent loop

A runtime agent should:

1. read relevant projections
2. open or use the appropriate channel
3. explain intended user-visible action
4. execute generated command tools
5. verify resulting projection state
6. post a concise channel update

See [Runtime MCP](/agents-in-cessy/runtime-mcp).

## Safety rule

Give agents the narrowest MCP authority that matches the job:

* read-only Design MCP for discovery and review
* full Design MCP for trusted design mutation
* Runtime MCP for published app operation
* API keys only for automation that cannot use OAuth

When a task crosses layers, make the handoff explicit.

> Use Design MCP for builder agents and Runtime MCP for published app agents without mixing design-time and runtime authority.

## Docs

- [Runtime MCP](https://docs.cessy.ai/agents-in-cessy/runtime-mcp.md): Use generated Runtime MCP tools for published app commands, projection and file reads, event operations, channels, users, DMs, and page refresh.
- [Agent UX Patterns](https://docs.cessy.ai/agents-in-cessy/agent-ux-patterns.md): Design Cessy agent workflows that are inspectable, scoped, useful, and safe for builders, operators, and end users.
- [Agent Docs Contract](https://docs.cessy.ai/agents-in-cessy/agent-docs-contract.md): Keep Cessy documentation useful to human builders and directly actionable for AI agents connected through Design MCP or Runtime MCP.