Project Scaffold Generator

Purpose

Generate a complete, customized agentic development environment tailored to your specific project needs through guided questions.

Size Limits (Enforced)

All generated artifacts MUST respect these limits:

Progressive Disclosure Pattern

Structure ALL generated content using Anthropic's 3-tier pattern:

Tier 1: METADATA (~100 tokens)
├── name + description in frontmatter
├── Loaded at startup for ALL installed skills
└── Must clearly indicate WHEN to trigger

Tier 2: INSTRUCTIONS (1-5K tokens)  
├── Full SKILL.md body
├── Loaded only when skill is triggered
└── Core workflow — NOT reference material

Tier 3: RESOURCES (on-demand)
├── Bundled files in skill directory
├── Loaded only when explicitly referenced
└── Templates, checklists, examples, patterns

Generation Rules:

• Never embed examples >20 lines in SKILL.md — use See: examples.md
• Never embed templates in SKILL.md — use See: templates/[name].template.[ext]
• Never embed reference material — use See: reference.md
• Always create templates/ subdirectory for reusable code

When to Use

• Starting a new development project
• Adding AI coding assistant capabilities to existing project
• Standardizing team's AI tooling approach
• Migrating from basic to advanced agentic features

Workflow

Phase 1: Project Discovery

Ask the user these questions (use multiple-choice when possible):

1. What type of project is this?

   • New greenfield project
   • Existing codebase enhancement
   • Migration or refactor
   • Prototype/POC
   • Other (specify)

2. What is your primary technology stack?

   • Frontend: [React/Vue/Angular/Svelte/Other]
   • Backend: [Node.js/Python/Java/Go/Ruby/.NET/Other]
   • Database: [PostgreSQL/MySQL/MongoDB/Redis/Other]
   • Infrastructure: [AWS/Azure/GCP/On-prem/Hybrid]

3. What languages will be used? (multi-select)

   • JavaScript/TypeScript
   • Python
   • Java/Kotlin
   • C#/.NET
   • Go
   • Rust
   • Other (specify)

4. Team size and structure?

   • Solo developer
   • Small team (2-5 people)
   • Medium team (6-15 people)
   • Large team (16+ people)
   • Open source project

5. What are your primary development workflows? (multi-select)

   • Unit/integration testing
   • End-to-end testing
   • Code review process
   • CI/CD pipelines
   • Local development
   • Pair programming
   • Documentation generation
   • Performance monitoring
   • Security scanning

6. Desired AI integration level?

   • Basic (code completion, simple suggestions)
   • Intermediate (workflow automation, guided tasks)
   • Advanced (autonomous agents, complex orchestration)

7. What external tools need integration? (multi-select)

   • Git/GitHub/GitLab
   • Jira/Linear/Asana
   • Slack/Discord
   • Cloud provider CLI
   • Database tools
   • API testing tools
   • Monitoring/logging services
   • Other (specify)

Phase 2: Generate Core Configuration

Based on answers, create:

1. .github/copilot-instructions.md

   • Project overview from answers
   • Technology stack guidelines
   • Team conventions (based on team size)

2. Language-Specific Instructions For each selected language, create .github/instructions/[language].instructions.md:

   • Language-specific best practices
   • Framework patterns
   • Common gotchas

3. Framework Instructions (if applicable) For major frameworks (React, Django, etc.):

   • Framework-specific patterns
   • Project structure conventions
   • Common patterns for this stack

Phase 3: Generate Recommended Skills

Based on workflows selected, generate skills in .github/skills/.

Required Skill Directory Structure:

.github/skills/[skill-name]/
├── SKILL.md              # Core workflow (MAX 500 lines)
├── checklist.md          # Validation checklist (optional)
├── patterns.md           # Code patterns if >50 lines
├── examples.md           # Extended examples if >20 lines
└── templates/            # Reusable templates
    └── [name].template.[ext]

Testing Workflow → Create testing/SKILL.md:

---
name: testing
description: Comprehensive testing workflow including unit, integration, and e2e tests. Use when adding tests or debugging test failures.
---

# Testing Skill {#testing-skill }

## Purpose {#purpose }
Streamline test creation and execution for [PROJECT_STACK]

## When to Use {#when-to-use }
- Writing new tests for features
- Debugging failing tests
- Running test suites
- Updating test coverage

## Instructions {#instructions }

1. **Identify test type needed**:
   - Unit: Single function/component isolation
   - Integration: Multiple components working together
   - E2E: Full user workflow

2. **Generate test scaffold**:
   [Language-specific test template]

3. **Run tests**:
   - Command: [project-specific command]
   - Watch mode: [if applicable]

4. **Validate coverage**:
   - Check coverage report
   - Identify gaps
   - Add missing tests

## Examples {#examples }
[Stack-specific examples]