Skip to main content

Droost: Guidelines

droost_guidelines Read-only

Module: Droost core

The playbook, served over MCP: version-stamped Drupal conventions plus deep-dive topics from every enabled module — including this site's methodology (structured vs freeform vs listing) and theming doctrine. Read the topic before building the thing.

Example call
{
    "tool": "droost_guidelines",
    "arguments": {
        "topic": "methodology"
    }
}
Example response — captured live on this site
{
    "success": true,
    "message": "Guideline topic: methodology.",
    "data": {
        "topic": "methodology",
        "content": "# The Droost site-building methodology — read this before building anything\n\nThis site is built the Droost way: **every visible surface renders through\nSingle Directory Components (SDC)**, and every change flows through Droost's\ndeterministic, validated tools. You (the agent) are the design brain; Droost\nis the executor. Follow this playbook and your output will be reviewable,\nre-runnable, and green.\n\n## The one decision that shapes everything\n\nFor each thing the user wants on screen, ask: **is it a freeform page or a\nstructured entity?**\n\n| It is… | Build it as… | Compose with… |\n|---|---|---|\n| A landing page, marketing page, one-off layout | A **Canvas page** (`canvas_page` entity, or any entity with a `component_tree` field) | `droost_canvas_tree_set` |\n| Repeatable content with fields (events, products, articles, people…) | A **content type** whose *view display* is composed from SDCs | `droost_display_compose` |\n| A list of those entities (inventory, search page, featured row) | A **View** rendering the bundle's SDC-composed teasers | `droost_views_compose` |\n\nDo **both** on most sites: content types hold the data and render through\nSDC-composed displays; Canvas pages present curated entry points. **Never**\nleave a c
… (truncated — run it yourself for the rest)

Returns curated Drupal development guidelines for AI agents. No arguments: the core conventions plus a list of available deep-dive topics. With "topic": that topic's full content. Read-only.

Input schema (JSON Schema)
{
    "type": "object",
    "properties": {
        "topic": {
            "type": "string",
            "description": "A topic machine name (from the listed topics), e.g. \"entity-api\", \"plugins\", \"mcp-tools\"."
        }
    }
}