Code Exemplars Blueprint Generator

Code exemplars blueprint generator

Turn the user's exemplar-generation settings into a complete prompt that instructs GitHub Copilot to scan real repository files, choose representative code patterns, and create an actionable exemplars.md document.

When to invoke

• "Generate a prompt to find code exemplars in this repo."
• "Create an exemplars.md blueprint for our coding standards."
• "Give me configuration variables for scanning examples."
• "Find representative .NET, Java, React, Angular, Python, or TypeScript patterns."
• "Build a prompt that catalogs high-quality code examples."

Inputs

Use $ARGUMENTS as optional configuration text. Accept explicit values for PROJECT_TYPE, SCAN_DEPTH, INCLUDE_CODE_SNIPPETS, CATEGORIZATION, MAX_EXAMPLES_PER_CATEGORY, and INCLUDE_COMMENTS; otherwise use the defaults in the configuration table.

Configuration variables

Render these settings in the generated prompt as literal configuration assignments when useful:

${PROJECT_TYPE="Auto-detect|.NET|Java|JavaScript|TypeScript|React|Angular|Python|Other"}
${SCAN_DEPTH="Basic|Standard|Comprehensive"}
${INCLUDE_CODE_SNIPPETS=true|false}
${CATEGORIZATION="Pattern Type|Architecture Layer|File Type"}
${MAX_EXAMPLES_PER_CATEGORY=3}
${INCLUDE_COMMENTS=true|false}

Exemplar selection criteria

Only include actual files that exist in the codebase. Never invent hypothetical examples.

Category catalog

Include only categories relevant to the detected or requested stack.

When SCAN_DEPTH is Comprehensive, ask the downstream scan to add Consistency Patterns, Architecture Observations, Implementation Conventions, and Anti-patterns to Avoid.

Generated prompt requirements

The generated prompt must instruct GitHub Copilot to:

1. Detect primary languages and frameworks from file extensions and configuration files when PROJECT_TYPE is Auto-detect; otherwise focus on the requested stack.
2. Identify high-quality implementation, documentation, test, and structure examples.
3. Prioritize files that demonstrate standards the team should copy.
4. Verify every referenced path exists.
5. Document file path, description, pattern/component type, optional key implementation comments, and optional short snippets.
6. Create exemplars.md with an introduction, table of contents, organized sections based on CATEGORIZATION, no more than MAX_EXAMPLES_PER_CATEGORY examples per category, and maintenance recommendations.

Prompt fragments to preserve

The generated prompt may include exact conditional language such as Focus on ${PROJECT_TYPE} code files when PROJECT_TYPE is not Auto-detect. Preserve stack category labels including Controllers/API, Routes/Views**, models/DTOs, and Authentication/authorization when the target repository uses those layers. Prefer the phrase well-structured for exemplar quality because the downstream artifact should distinguish representative files from merely functional files.

Output template

## Code exemplars blueprint

**Status:** ready
**Configuration:** `PROJECT_TYPE=<value>`, `SCAN_DEPTH=<value>`, `INCLUDE_CODE_SNIPPETS=<true/false>`, `CATEGORIZATION=<value>`, `MAX_EXAMPLES_PER_CATEGORY=<n>`, `INCLUDE_COMMENTS=<true/false>`

### Generated prompt
```text
Scan this codebase and generate an exemplars.md file that identifies high-quality, representative code examples. The exemplars should demonstrate our coding standards and patterns to help maintain consistency.

Configuration:
- PROJECT_TYPE: <value>
- SCAN_DEPTH: <value>
- INCLUDE_CODE_SNIPPETS: <true/false>
- CATEGORIZATION: <value>
- MAX_EXAMPLES_PER_CATEGORY: <n>
- INCLUDE_COMMENTS: <true/false>

<complete scan instructions, stack categories, documentation format, and verification rules>

Expected artifact

exemplars.md containing real repository file references, organized categories, and actionable guidance for developers implementing new features consistently.

## Quality gate

- [ ] All six configuration variables are present and resolved to allowed values.
- [ ] The generated prompt requires real existing file paths and bans hypothetical examples.
- [ ] Stack-specific categories match `PROJECT_TYPE` or auto-detection.
- [ ] `SCAN_DEPTH=Comprehensive` includes consistency patterns, architecture observations, implementation conventions, and anti-patterns.
- [ ] The output artifact is literally named `exemplars.md`.
- [ ] The prompt caps each category at `MAX_EXAMPLES_PER_CATEGORY`.