
GitHub Copilot app for Beginners: How to teach Copilot about your codebase
Beginner-oriented GitHub blog guide to giving GitHub Copilot project context through four mechanisms: repository custom instructions (.github/copilot-instructions.md and path-specific .github/instructions files with applyTo), agent skills (SKILL.md folders under .github/skills), custom agents (.github/agents profiles), and MCP servers such as Playwright. It explains when to use each and how they combine, using examples like a quality-checks skill and an accessibility-review agent. Useful as a concrete reference for the repo layout Copilot reads, with a pointer to the community Awesome Copilot collection.
What it claims
Copilot reads project context from four places in a repository. Each one covers a different gap:
- Instructions cover conventions you keep re-explaining.
- Skills cover multi-step procedures you repeat.
- Custom agents cover recurring large tasks that need a dedicated role.
- MCP servers cover tools and data outside the repo.
The page treats these as layers you combine, not alternatives.
How it works
Instructions stack rather than override. When Copilot works on a file that matches a path-specific file's applyTo, it can use both .github/copilot-instructions.md and the path-specific guidance. Keep broad rules in the repo-wide file and put framework- or file-type-specific rules in .github/instructions.
Skill selection is driven by the description. Copilot uses the description in SKILL.md to decide when to load a skill, so that text determines whether the skill ever fires. A skill folder can also hold scripts and other resources besides SKILL.md. The page's suggested content for a quality-checks skill:
- when Copilot should use the skill
- which scripts run the project's tests and linting
- what each script checks
- how to troubleshoot common failures
The stated payoff is that Copilot runs your actual test and lint commands instead of guessing them.
Agent profiles in .github/agents define the agent's purpose, instructions, tools and relevant project context. Use one when the work spans many files and draws on several tools or sources. The page's example is an accessibility review covering components, templates, styles and interactions.
MCP via Playwright. Browse MCP servers in the Copilot app or add a custom one. Once Playwright is configured, you invoke it explicitly in the prompt:
Use the Playwright MCP server to test the functionality manually.Copilot then opens the app in a browser and interacts with it. This checks UI behaviour beyond what automated checks cover.
How to use it
The recommended entry point is a single step: pick one convention you keep explaining, add it to .github/copilot-instructions.md, and watch how Copilot applies it on the next task. For skills and agents, take ready-made ones from Awesome Copilot, review what they contain, and adapt them to your repo's tools.
Limits
This is a beginner walkthrough with no file-format reference:
- It shows no example
applyTosyntax, noSKILL.mdfrontmatter, no agent profile schema and no MCP configuration file. - It gives no evidence of how reliably Copilot picks up skills or path instructions.
- The guidance applies to "the Copilot app". The page does not say which other Copilot surfaces read these files.
Organisations named here
Technologies named here
- GitHub Copilot app
- Custom instructions
- Path-specific instruction files
- Custom agents
- Model Context Protocol servers
- Playwright MCP server
- Awesome Copilot
- ReviewBench
- GitHub Copilot
- Model Context Protocol
- Playwright
- GitHub
Standards and laws named here
Techniques named here
- Repository custom instructions
- Custom agents
- MCP server integration for verification
Related in coding-agents
