AgentSkillsCN

copilot-spark

借助 GitHub Copilot Spark,以自然语言构建应用程序。特别适合快速原型开发、向非编程人员传授编程知识,以及在正式投入全面实施前探索各类创意。可通过“Spark”“原型”“应用构建器”“无代码”“低代码”“自然语言应用”“教授编程”“入职培训”等关键词触发。

SKILL.md
--- frontmatter
name: copilot-spark
description: Build apps with natural language using GitHub Copilot Spark. Perfect for rapid prototyping, teaching non-coders, and exploring ideas before committing to full implementation. Triggers on spark, prototype, app builder, no-code, low-code, natural language app, teach coding, onboarding.

GitHub Copilot Spark Skill

AI-powered app builder for rapid prototyping and teaching

GitHub Copilot Spark transforms natural language descriptions into working web applications. Use it to prototype ideas quickly, teach programming concepts, or bridge the gap between idea and implementation.

Prerequisites

  • GitHub Account: With Copilot access
  • Spark Access: Visit spark.github.com or use gh spark CLI
  • Browser: Modern browser for the Spark editor

Quick Reference

Core Workflows

WorkflowDescriptionBest For
PrototypeDescribe app → Get working codeRapid iteration, proof of concept
TeachExplain concepts through buildingOnboarding, learning
ExploreTry different approaches quicklyArchitecture decisions
BridgeMove from Spark to production codeHandoff to development

Key Commands

ActionMethod
Create new appDescribe in natural language
Modify app"Change the button to blue"
Add feature"Add a dark mode toggle"
Export codeDownload or copy generated code
Share prototypeGenerate shareable link

Use Cases

1. Rapid Prototyping

code
User: "Create a todo list app with categories and due dates"

Spark generates:
- React components for todo items
- Category filtering
- Date picker for due dates
- Local storage persistence

2. Teaching Non-Coders

code
User: "Show me how a login form works"

Spark creates:
- Email/password form
- Basic validation
- Submit handling
- Explains each part

3. Architecture Exploration

code
User: "Build a dashboard with charts showing sales data"

Spark provides:
- Chart component options
- Data structure suggestions
- Layout alternatives

Spark to Production Workflow

Step 1: Prototype in Spark

code
Describe your idea → Iterate until satisfied → Export code

Step 2: Review Generated Code

bash
# Spark exports typically include:
- src/components/     # React components
- src/styles/        # CSS/styling
- src/utils/         # Helper functions
- package.json       # Dependencies

Step 3: Integrate with Production Codebase

bash
# Copy relevant components
cp -r spark-export/src/components/* ./src/components/spark-prototype/

# Review and refactor for production standards
# - Add TypeScript types
# - Add error handling
# - Add tests
# - Follow project conventions

Step 4: Production Hardening Checklist

  • Add TypeScript types/interfaces
  • Implement proper error handling
  • Add unit and integration tests
  • Apply project styling conventions
  • Add accessibility attributes
  • Implement proper state management
  • Add loading and error states
  • Security review (input validation, XSS prevention)

Best Practices

Effective Prompts

✅ Good Prompts❌ Avoid
"Todo app with drag-and-drop reordering""Make an app"
"Dashboard showing user activity metrics""Analytics thing"
"Form with email validation and submit""Some inputs"
"Card grid with hover effects""Display stuff"

Iteration Tips

  1. Start simple: Begin with core functionality
  2. Iterate incrementally: Add features one at a time
  3. Be specific: "Blue button" vs "styled button"
  4. Reference examples: "Like Twitter's compose box"

Teaching Approach

  1. Show, don't tell: Let Spark generate, then explain
  2. Break it down: Ask for one concept at a time
  3. Compare approaches: "Show me two ways to do this"
  4. Explain the why: Ask Spark to comment the code

Integration with This Repository

Using Spark Prototypes

bash
# 1. Create prototype in Spark
# 2. Export to local directory
# 3. Use the bridge prompt to integrate

/spark-bridge --source ./spark-export --target ./src/features/new-feature

Related Resources

ResourcePurpose
@spark-prototyperAgent for guided prototyping
/spark-prototypeQuick prototype prompt
/spark-teachTeaching/onboarding prompt
reference.mdDetailed command reference

Limitations

  • Complexity: Best for simple to medium complexity apps
  • Backend: Limited backend/API generation
  • State: Basic state management only
  • Testing: No test generation
  • Types: Limited TypeScript support

When to Use Spark vs. Traditional Development

Use Spark WhenUse Traditional Dev When
Exploring ideasProduction code
Quick demosComplex logic
Teaching conceptsTeam collaboration
UI prototypingBackend services
Client presentationsSecurity-critical

Related Documentation