astro

Astro

Astro is an all-in-one web framework for building content-driven websites. It uses island architecture and server-first design to minimize client-side JavaScript. It supports UI frameworks like React, Vue, Svelte, Preact, and Solid via integrations, and is powered by Vite.

Research Process

To answer questions about Astro, spawn a sub-agent to conduct thorough documentation research.

Launch a Research Sub-Agent

Use the Task tool to spawn a sub-agent dedicated to Astro documentation research. Do NOT specify a model (use the default model for comprehensive research quality).

Critical: Provide the sub-agent with a highly detailed prompt that includes:

1. Goal - What you are ultimately trying to accomplish
2. Context - Why you need this information and how it fits into the larger task
3. Specific Questions - Exactly what information you need answered
4. Output Requirements - The format and level of detail needed in the response

Task Tool Configuration

Task tool parameters:
- description: "Research Astro docs"
- subagent_type: "generalPurpose"
- model: (DO NOT SPECIFY - use default for thorough research)
- readonly: true
- prompt: (see template below)

Prompt Template for Sub-Agent

Structure your prompt to the research sub-agent as follows:

You are researching Astro framework documentation to help with a specific task.

## Your Research Goal
[Describe exactly what you need to accomplish with this information]

## Context
[Explain why you need this information and how it will be used]

## Specific Questions to Answer
[List the specific questions that need to be answered]

## Research Process

1. First, fetch the documentation index to understand what's available:
   URL: https://docs.astro.build/llms.txt

   Astro provides curated documentation sets as separate files. Choose the
   most relevant one(s) based on the questions above:

   **General documentation (pick ONE based on needs):**
   - https://docs.astro.build/llms-small.txt - Abridged docs (compact, non-essential content removed). Use for quick lookups and common questions.
   - https://docs.astro.build/llms-full.txt - Complete docs (everything). Use when you need thorough or obscure information.

   **Topic-specific documentation (fetch alongside or instead of general docs):**
   - https://docs.astro.build/_llms-txt/api-reference.txt - API reference (terse, structured API descriptions)
   - https://docs.astro.build/_llms-txt/how-to-recipes.txt - How-to recipes (guided examples for adding features)
   - https://docs.astro.build/_llms-txt/build-a-blog-tutorial.txt - Blog tutorial (step-by-step beginner guide)
   - https://docs.astro.build/_llms-txt/deployment-guides.txt - Deployment guides (deploying to various services)
   - https://docs.astro.build/_llms-txt/cms-guides.txt - CMS guides (integrating content management systems)
   - https://docs.astro.build/_llms-txt/backend-services.txt - Backend services (Firebase, Sentry, Supabase, etc.)
   - https://docs.astro.build/_llms-txt/migration-guides.txt - Migration guides (migrating from other tools)
   - https://docs.astro.build/_llms-txt/additional-guides.txt - Additional guides (e-commerce, auth, testing, DAM)

2. Selection strategy:
   - For API/config questions → fetch api-reference.txt
   - For "how do I add X" questions → fetch how-to-recipes.txt
   - For deployment questions → fetch deployment-guides.txt
   - For CMS integration → fetch cms-guides.txt
   - For backend/third-party services → fetch backend-services.txt
   - For migrating from another framework → fetch migration-guides.txt
   - For e-commerce, auth, testing → fetch additional-guides.txt
   - For beginner/tutorial questions → fetch build-a-blog-tutorial.txt
   - For general or broad questions → fetch llms-small.txt (or llms-full.txt if depth needed)
   - When unsure → start with llms-small.txt, then fetch topic-specific files as needed

3. These are large text files containing full documentation content.
   Search through the fetched content for relevant sections.

4. Fetch multiple files in parallel using WebFetch when the question
   spans multiple topics.

## Required Output
[Specify exactly what format and content you need returned]

Provide a comprehensive response with all findings, including relevant code examples, configuration snippets, and step-by-step instructions where applicable.