Repository Analysis & Instruction System Generator
Overview
You are an expert AI system architect tasked with analyzing the current repository and creating a comprehensive instruction system with specialized persona prompt files that work across both Claude Code and GitHub Copilot.
Primary Objectives
- Analyze Repository: Examine the current codebase, architecture, dependencies, and project structure
- Generate CLAUDE.md: Create contextually appropriate instructions based on the code analysis (with both @import and markdown link syntax for cross-compatibility)
- Generate .claude/rules/: Create path-specific rules for different file types found in the repo
- Generate .github/instructions/: Create GitHub Copilot equivalents of the rules
- Create Skills: Build specialized skill files for different workflows
Required Analysis Steps
1. Repository Deep Dive
Perform a comprehensive analysis of:
- Language & Framework Detection: Identify primary programming languages, frameworks, libraries, and dependencies
- Architecture Patterns: Analyze code structure, design patterns, architectural decisions
- Project Type: Determine if it's a web app, API, mobile app, desktop app, library, etc.
- Build System: Identify build tools, package managers, CI/CD configurations
- Code Quality: Assess testing frameworks, linting rules, code organization
2. Generate Files
Create the following cross-compatible file structure:
CLAUDE.md # Main instructions (both tools)
.claude/
prompt-snippets/ # Shared prompt text
rules/ # Claude Code path-specific rules
skills/ # Shared skills
agents/ # Claude Code agents
.github/
copilot-instructions.md # GitHub Copilot instructions
instructions/ # GitHub Copilot path-specific rules
Execution
- Analyze the repository thoroughly
- Generate all files following the cross-compatibility patterns
- Ensure prompt snippets are the single source of truth
- Test that references between files are correct
Source: raffertyuy/github-copilot-prompts — distributed by TomeVault.