# Code Exemplars Blueprint Generator > Generate a configurable prompt blueprint for scanning a codebase and producing an exemplars.md catalog of high-quality, real code examples. Use this skill when the user asks for code exemplars, exemplar blueprint prompts, coding standard examples, repository pattern catalogs, or configuration variables for exemplar generation. - Skill: `paulasilvatech/code-exemplars-blueprint-generator` (Agent Skill) - Install (CLI): `npx skillmds@latest add paulasilvatech/code-exemplars-blueprint-generator` - Raw SKILL.md: https://api.skillmd.com/api/skills/paulasilvatech/code-exemplars-blueprint-generator/raw - Safety review: PASS (external: skill-scanner PASS, skillspector CAUTION) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: AI & ML - Author: paulasilvatech (https://skillmd.com/u/paulasilvatech) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/paulasilvatech/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 | Variable | Allowed values | Default | Controls | | --- | --- | --- | --- | | `PROJECT_TYPE` | `Auto-detect`, `.NET`, `Java`, `JavaScript`, `TypeScript`, `React`, `Angular`, `Python`, `Other` | `Auto-detect` | Which languages and frameworks the generated prompt emphasizes. | | `SCAN_DEPTH` | `Basic`, `Standard`, `Comprehensive` | `Standard` | How much project-wide convention and architecture analysis to request. | | `INCLUDE_CODE_SNIPPETS` | `true`, `false` | `true` | Whether `exemplars.md` includes short snippets as well as file paths. | | `CATEGORIZATION` | `Pattern Type`, `Architecture Layer`, `File Type` | `Pattern Type` | How examples are grouped. | | `MAX_EXAMPLES_PER_CATEGORY` | Positive integer | `3` | Maximum exemplars in each category. | | `INCLUDE_COMMENTS` | `true`, `false` | `true` | Whether each exemplar explains key implementation details and principles. | Render these settings in the generated prompt as literal configuration assignments when useful: ```text ${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. | Criterion | What to look for | Reject when | | --- | --- | --- | | Readability | Clear naming, small units, visible intent | Clever but opaque implementation. | | Documentation | Helpful comments, docstrings, or README-backed usage | Comments repeat syntax without explaining why. | | Error handling | Explicit validation, typed errors, recoverable failure paths | Exceptions swallowed or converted to vague messages. | | Architecture fit | Separation of concerns, dependency direction, conventional boundaries | Feature code bypasses layers or central conventions. | | Tests | Representative unit, integration, or component tests | Tests only assert imports or snapshots with no behavior. | | Maintainability | Low duplication, simple control flow, standard patterns | Code smell is present even if the file is common. | ## Category catalog Include only categories relevant to the detected or requested stack. | Stack | Candidate categories | | --- | --- | | `.NET` | Domain Models, Repository Implementations, Service Layer Components, Controller Patterns, Dependency Injection Usage, Middleware Components, Unit Test Patterns. | | Frontend (`JavaScript`, `TypeScript`, `React`, `Angular`) | Component Structure, State Management, API Integration, Form Handling, Routing Implementation, UI Components, Unit Test Examples. | | `Java` | Entity Classes, Service Implementations, Repository Patterns, Controller/Resource Classes, Configuration Classes, Unit Tests. | | `Python` | Class Definitions, API Routes/Views, Data Models, Service Functions, Utility Modules, Test Cases. | | Architecture layer grouping | Presentation Layer, Business Logic Layer, Data Access Layer, Cross-Cutting Concerns such as logging, error handling, authentication/authorization, and validation. | 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 ```markdown ## Code exemplars blueprint **Status:** ready **Configuration:** `PROJECT_TYPE=`, `SCAN_DEPTH=`, `INCLUDE_CODE_SNIPPETS=`, `CATEGORIZATION=`, `MAX_EXAMPLES_PER_CATEGORY=`, `INCLUDE_COMMENTS=` ### 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: - SCAN_DEPTH: - INCLUDE_CODE_SNIPPETS: - CATEGORIZATION: - MAX_EXAMPLES_PER_CATEGORY: - INCLUDE_COMMENTS: ``` ### 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`.