测试助手
- 作者仓库星标 59,805
- 作者更新于 实时读取
- 作者仓库 astro
- 领域
- AI 智能
- 兼容 Agent
-
- Claude Code
- Cursor
- Cline
- Codex
- Windsurf
- Gemini CLI
- +20
- 信任分
- 88 / 100 · 社区维护
- 作者 / 版本 / 许可
- @withastro · 未声明 license
- Token 消耗评级
- 低消耗
- 接入复杂程度
- 需简单配置
- 是否需要外部 API Key
- 不需要
- 兼容的系统
- macOS · Linux · Windows
- 底层运行要求
- Node.js
- 文件与系统权限
-
- 只读
- 允许写入 / 修改
- Shell 执行
- 网络行为
- 仅限本地
- 安装命令数
- 26 条
档案由构建时根据 SKILL.md 与安装命令自动衍生,可能与作者实际意图存在差异。
需要注意: 未限定 allowed-tools,默认拥有全部工具权限。
---
name: astro-developer
description: Comprehensive guide for developing in the Astro monorepo. Covers architecture, debugging, testin…
category: AI 智能
runtime: Node.js
---
# astro-developer 输出预览
## PART A: 任务判断
- 适用问题:提示词、Agent 工作流、模型评估或自动化推理。
- 输入要求:目标材料、限制条件、期望输出和验收方式。
- 证据边界:围绕“Quick Decision Matrix / Critical Warnings / Quick Command Reference”读取原文规则,不把推断写成作者承诺。
## PART B: 执行结果
- **01** 任务判断:确认你的需求是否属于提示词、Agent 工作流、模型评估或自动化推理,并标出输入、限制和预期结果。
- **02** 执行计划:优先按“Quick Decision Matrix / Critical Warnings / Quick Command Reference”拆成步骤,说明每一步会读取什么、修改什么、产出什么。
- **03** 交付结果:给出可复制的命令、文件改动、检查清单或内容草稿,并说明如何继续迭代。
- **04** 风险边界:结合 读取文件、写入/修改文件、执行终端命令、主要在本地完成、通常不需要额外 API Key 给出执行前确认项。
## Running Rules
- 读取文件、写入/修改文件、执行终端命令;主要在本地完成;通常不需要额外 API Key。
- 先小样例验证,再放大到真实任务。
- 交付时同时给结果、检查口径和下一步迭代建议。 原文没有稳定的斜杠命令要求。安装验证后通常全局生效,直接在对话里点名这个 Skill 并描述任务即可。
告诉 Agent 目标文件或材料、期望结果、不可改范围、是否允许联网或执行命令。本 Skill 的权限画像是:读取文件、写入/修改文件、执行终端命令。
先用一个小任务确认它会围绕“Quick Decision Matrix / Critical Warnings / Quick Command Reference”工作;涉及文件或命令时,先看 diff、日志、预览或测试结果。
检查最终产物是否包含明确结果、必要证据和下一步动作;如果输出泛泛而谈,就补充输入、边界和验收标准后重跑。
---
name: astro-developer
description: Comprehensive guide for developing in the Astro monorepo. Covers architecture, debugging, testin…
category: AI 智能
source: withastro/astro
---
# astro-developer
## 什么时候使用
- 用于组织测试、定位失败并形成修复闭环 适合处理AI Agent、提示词、模型评估与自动化推理,核心价值是把输入、判断、执行、验证和交付边界固定下来,避免 Agent 泛泛回答。 把任务拆成可执行、可检查、可继续迭代的步骤;通常不需要额外…
- 面向提示词、Agent 工作流、模型评估或自动化推理,优先处理能明确输入、步骤和验收标准的工作。
## 需要提供什么
- 目标材料、目录范围、期望结果和不可改动内容。
- 是否允许联网、执行命令、读写文件或调用外部服务。
## 执行规则
- 围绕「Quick Decision Matrix / Critical Warnings / Quick Command Reference」组织步骤,不把推断写成作者事实。
- 读取文件、写入/修改文件、执行终端命令;主要在本地完成;通常不需要额外 API Key。
- 先跑小样例,确认结果可检查后再扩大任务范围。
## 输出要求
- 给出最终产物、关键证据、验证方式和下一步动作。
- 信息不足时标记 unknown,不编造命令、平台或依赖。 作者原文负责流程事实;仓库文件负责来源和命令;流狐只补充适用场景、限制和质量判断。
skill "astro-developer" {
输入层 -> 用户目标 + 目标文件 + 禁止范围 + 验收标准
上下文层 -> Quick Decision Matrix / Critical Warnings / Quick Command Reference
规则层 -> SKILL.md 触发条件 / 执行顺序 / 输出格式
运行层 -> Node.js | 读取文件、写入/修改文件、执行终端命令 | 主要在本地完成
安全层 -> 通常不需要额外 API Key + 小任务验证 + diff / 日志复核
输出层 -> 可复制结果 + 检查清单 + 下一步迭代
} Astro Developer Skill
Context-loading skill for AI agents and developers working in the Astro monorepo. Loads relevant documentation based on your task.
Quick Decision Matrix
What are you doing? → Read these files:
| Task | Primary Docs | Supporting Docs |
|---|---|---|
| Adding a core feature | architecture.md, constraints.md | testing.md |
| Fixing a bug | debugging.md | architecture.md |
| Writing/fixing tests | testing.md | constraints.md |
| Creating an integration | Explore packages/integrations/ for examples |
testing.md |
| Understanding architecture | architecture.md | - |
| Dealing with errors | debugging.md, constraints.md | testing.md |
| Understanding constraints | constraints.md | architecture.md |
Critical Warnings
Before you start, be aware of these common pitfalls:
- Prefer Unit Tests: Write unit-testable code by default. Use integration tests only when necessary → testing.md
- Node.js API Restrictions: Cannot use Node.js APIs in
runtime/code → constraints.md - Test Isolation: Must set unique
outDirfor each integration test → testing.md - Runtime Boundaries: Core vs Vite vs Browser execution contexts → architecture.md
- Prerelease Mode: Changesets target
origin/nextbranch (check.changeset/config.json)
Quick Command Reference
# Development
pnpm install # Install (root only)
pnpm run build # Build all packages
pnpm run dev # Watch mode
pnpm run lint # Lint codebase
# Testing
pnpm -C packages/astro exec astro-scripts test "test/**/*.test.js" # All tests
pnpm -C packages/astro exec astro-scripts test -m "pattern" # Filter tests
pnpm run test:e2e # E2E tests
node --test test/file.test.js # Single test
# Examples
pnpm --filter @example/minimal run dev # Run example
# Changesets
pnpm exec changeset --empty # Create changeset, no interactive mode
Key File Paths
packages/astro/src/
├── core/ # Node.js execution context (build/dev commands)
├── runtime/
│ ├── server/ # Vite SSR execution context
│ └── client/ # Browser execution context
├── virtual-modules/ # Virtual module entry points
├── content/ # Content layer system
├── vite-plugin-*/ # Vite plugins
└── types/ # Centralized TypeScript types
packages/integrations/ # Official integrations
examples/ # Test your changes here
test/fixtures/ # Test fixtures
Note: Error stack traces in node_modules/ map to source in packages/. See architecture.md for details.
Usage
This skill loads relevant context—it doesn't orchestrate workflows. After loading appropriate docs:
- Read the recommended files for your task
- Apply the patterns and constraints described
- Use the commands and file paths provided
- Search docs for error messages if you encounter issues
Architecture Quick Summary
Three Execution Contexts:
- core/ → Node.js, build/dev commands, avoid Node APIs except in Vite plugins
- runtime/server/ → Vite SSR, CANNOT use Node APIs
- runtime/client/ → Browser, CANNOT use Node APIs at all
Five Pipeline Types:
- RunnablePipeline →
astro devwith Vite loader system - NonRunnablePipeline →
astro devwithout runtime module loading (Cloudflare adapter) - BuildPipeline →
astro build+ prerendering - AppPipeline → Production serverless/SSR
- ContainerPipeline → Container API
See architecture.md for complete details.
Testing Quick Summary
Philosophy: Prefer unit tests over integration tests. Write unit-testable code by default.
Unit tests (fast, preferred):
- Test pure functions and business logic
- Extract business logic from infrastructure
- Use dependency injection
Integration tests (slow, use sparingly):
- Only for features that cannot be unit tested (virtual modules, full build pipeline)
- Always set unique
outDirto avoid cache pollution
See testing.md for complete patterns and examples.
When NOT to Use This Skill
- Bug triage: Use the
triageskill instead - GitHub Actions analysis: Use the
analyze-github-action-logsskill - Simple questions: Just ask directly, don't load this skill
Related Documentation
- Root: /AGENTS.md
- Root: /CONTRIBUTING.md
- Astro docs: https://docs.astro.build/llms.txt
- Package: packages/astro/src/core/README.md
- Build plugins: packages/astro/src/core/build/plugins/README.md
先判断是否适合
作者设计意图
作者的方法与取舍
边界和复核