This repository contains a collection of GitHub Copilot customization files — agents, skills, prompts, and workspace instructions — designed to enhance AI-assisted development for Java and .NET projects.

.github/
├── copilot-instructions.md   # Workspace-level Copilot instructions
├── agents/                   # Custom agent mode definitions
│   ├── codereview.agent.md
│   ├── CSharpExpert.agent.md
│   ├── JavaExpert.agent.md
│   ├── documentation_generator.md
│   ├── feature_implementer.md
│   └── unit_test_generator.md
├── prompts/                  # Reusable prompt files
│   ├── changelog.prompt.md
│   └── create-readme.prompt.md
└── skills/                   # Skill files loaded on demand
    ├── csharp-async/
    ├── dotnet-best-practices/
    ├── dotnet-timezone/
    ├── java-async/
    ├── java-best-practices/
    ├── java-timezone/
    └── prompt-master/

File: .github/copilot-instructions.md

Copilot instructions are automatically loaded by GitHub Copilot for every conversation in the workspace. They define the project's technology stack, constraints, and coding conventions so Copilot stays aligned with the project without requiring repeated context.

The current instructions target a Java 21 Vaadin Web App called PersonalFinanceApp:

• Stack: Java 21, Spring Boot 3, Spring Data JPA (Hibernate), SQLite
• Package conventions: model, data, service
• Testing: JUnit 5
• Money: BigDecimal
• Async: CompletableFuture / Spring @Async
• Constraints: no auth, no charts, no CQRS, no Clean Architecture layers

Place a copilot-instructions.md file inside .github/ at the root of your repository. VS Code with the GitHub Copilot extension will automatically pick it up. No additional configuration is required.

Directory: .github/agents/

Agent files (.agent.md) define custom Copilot agent modes — specialized personas with tailored instructions, tool permissions, and model choices. Each agent is available in VS Code's Copilot Chat via the agent picker.

1. Open Copilot Chat in VS Code.
2. Click the agent picker (the @ icon or the model/agent selector in the chat input).
3. Select the desired agent from the list.
4. Start chatting — the agent's instructions and tool permissions are applied automatically.

The Feature Implementer agent is particularly powerful: it invokes the Unit Test Generator and Documentation Generator as subagents automatically after implementing a feature.

Directory: .github/skills/

Skills are instruction files that are loaded on demand by an agent when a relevant task is detected. Unlike copilot-instructions.md (always active), skills are only injected into context when needed, keeping token usage efficient.

Each skill is a SKILL.md file inside its own subdirectory. To make a skill available to Copilot agents, reference it in the agent's system prompt or the workspace copilot-instructions.md using the skill's path:

<!-- Inside an agent .md file or copilot-instructions.md -->
Read and apply the skill at `.github/skills/java-async/SKILL.md` when the user asks about async Java code.

Skills can also be invoked explicitly in chat:

Apply the java-best-practices skill to the selected code.

Directory: .github/prompts/

Prompt files (.prompt.md) are reusable, standalone prompts that can be invoked directly from Copilot Chat. They are useful for repeatable tasks like generating changelogs or README files.

Updates the project's changelog based on the git commit history.

• Model: Claude Sonnet 4.6
• Use when: You want to generate or update CHANGELOG.md after merging features or fixes.

How to invoke:

/changelog

Or open the command palette and select Copilot: Run Prompt, then choose changelog.

Generates a comprehensive, well-structured README.md for the project.

• Mode: Agent (explores the full workspace)
• Use when: Starting a new project or refreshing outdated documentation.

How to invoke:

/create-readme

The prompt instructs Copilot to review the entire workspace, take inspiration from real-world open-source README examples, and produce a concise, GFM-formatted README without over-using emojis.

┌─────────────────────────────────────────────────┐
│            copilot-instructions.md              │
│   (always active — project context & rules)     │
└────────────────────┬────────────────────────────┘
                     │
        ┌────────────┴────────────┐
        │                         │
   ┌────▼─────┐             ┌─────▼──────┐
   │  Agents  │             │   Prompts  │
   │ (modes)  │             │ (tasks)    │
   └────┬─────┘             └────────────┘
        │
   ┌────▼─────┐
   │  Skills  │
   │(on demand│
   │ context) │
   └──────────┘

1. copilot-instructions.md sets the baseline for every Copilot interaction in the workspace.
2. Agents provide specialized personas for focused tasks (code review, feature implementation, etc.).
3. Skills inject domain-specific knowledge (async patterns, timezone handling, best practices) when needed.
4. Prompts automate repeatable multi-step workflows (changelog, README generation).

Create .github/agents/my-agent.agent.md with YAML frontmatter:

---
name: My Agent
description: What this agent does
tools: ['read', 'edit', 'search']
model: Claude Sonnet 4.6 (copilot)
---

Your agent instructions here.

Create .github/skills/my-skill/SKILL.md:

---
name: my-skill
description: When to apply this skill
---

Your skill content here.

Create .github/prompts/my-task.prompt.md:

---
description: What this prompt does
---

Your prompt instructions here.