Complete Reference & Links

Comprehensive overview of Copilot CLI — modes of use, use cases, security considerations, model usage, and the ACP protocol. The best starting point.

Quickstart walkthrough — install, launch, authenticate, and start coding with Copilot CLI in minutes.

Installation instructions for all platforms — curl script, Homebrew, WinGet, npm. Includes version pinning and custom install paths.

OAuth device flow, Personal Access Tokens (PAT), GH_TOKEN and GITHUB_TOKEN environment variables, multi-account support.

Trusted directories, allowed tools, path and URL permissions, config.json settings, themes, and terminal configuration.

Diagnose authentication failures — token issues, proxy configs, enterprise SSO, and common error messages.

Complete usage guide — plan mode, autopilot, custom instructions, custom agents, skills, MCP servers, delegating to coding agent, session management, and more.

Expert tips — custom instructions, model selection, plan mode workflow, infinite sessions, delegating work, TDD, codebase onboarding, multi-repo patterns, and advanced CLI patterns.

Complete reference for all CLI commands, slash commands, keyboard shortcuts, command-line options, and environment variables. The definitive cheat sheet.

Side-by-side comparison of all customization features — custom instructions, prompt files, custom agents, subagents, skills, MCP servers. Includes IDE support matrix.

Complete plugin system reference — plugin.json schema, marketplace.json schema, install formats, file locations, loading order, precedence rules, and deduplication.

All hook types with input/output JSON schemas, script examples (Bash + PowerShell), advanced patterns — compliance audit trails, cost tracking, code quality enforcement, notifications.

Agent profile Markdown format, YAML frontmatter schema, tool restrictions, MCP server connections, infer settings, and naming conventions.

Detailed breakdown of which types of custom instructions are supported in each IDE and surface — VS Code, Visual Studio, JetBrains, Eclipse, Xcode, CLI, Coding Agent.

Full feature-by-IDE matrix — which features are supported (✓), under preview (P), or not supported (✗) in VS Code, Visual Studio, JetBrains, Eclipse, Xcode, Neovim.

Agent Client Protocol server documentation — use Copilot CLI as an agent in third-party tools, IDEs, or automation systems via the open ACP standard.

Overview of all customization options available for Copilot CLI — the best starting point for understanding what can be configured and when to use each option.

Repository-wide (.github/copilot-instructions.md), path-specific (.github/instructions/**/*.instructions.md), agent instructions (AGENTS.md, CLAUDE.md, GEMINI.md), local (~/.copilot/copilot-instructions.md), and COPILOT_CUSTOM_INSTRUCTIONS_DIRS.

Create SKILL.md with YAML frontmatter, project vs personal skills, skill directory structure, referencing scripts and resources, /skills commands, auto-detection vs manual invocation.

Add MCP servers with /mcp add, mcp-config.json format, GitHub MCP server (built-in), session-only servers with --additional-mcp-config, enabling/disabling servers.

Write .agent.md files with YAML frontmatter, define tools, MCP connections, infer settings. User-level (~/.copilot/agents/), repo-level (.github/agents/), org-level (.github-private/agents/).

Configure hooks in .github/hooks/*.json — sessionStart/End, preToolUse/postToolUse, userPromptSubmitted, agentStop, subagentStop, errorOccurred. Approve/deny tools programmatically.

Plugin directory structure, plugin.json manifest, bundling agents + skills + hooks + MCP configs, local testing with copilot plugin install, iterating on development.

Browse marketplaces, install from GitHub repos or local paths, plugin update, list, enable/disable, uninstall. Multiple install specification formats.

Build a marketplace.json registry, host on GitHub or local filesystem, share plugins with teams. Users add with /plugin marketplace add org/repo.

Deep comparison of custom instructions vs skills vs tools vs MCP servers vs hooks vs subagents vs custom agents vs plugins — when to use what, with decision table.

Let Copilot work autonomously until a task is complete. Permissions, --max-autopilot-continues, comparison with --allow-all and --no-ask-user, typical workflow, programmatic usage.

Parallel subagent execution — speed up multi-step implementation plans by running multiple subagents simultaneously. The /fleet command details.

How to use the /fleet command to parallelize multi-step plans for faster completion of large tasks.

Integrate Copilot CLI into CI/CD pipelines — run copilot programmatically in GitHub Actions workflows for automated code generation, review, and testing.

Enterprise administration — control Copilot CLI usage, policies, permissions, and MCP server restrictions at the organization and enterprise level.

Conceptual overview of the agent skills system — open standard (agentskills spec), project vs personal skills, how Copilot decides when to use a skill.

Conceptual overview of hooks — types (sessionStart, preToolUse, postToolUse, etc.), configuration format, JSON input/output, performance and security considerations.

Conceptual overview of custom agents — what they are, how they differ from subagents, agent profiles, auto-inference, tool restrictions, and scoping.

Persistent understanding of your repository — how Copilot stores and uses "memories" about coding conventions, patterns, and preferences across sessions.

What MCP is, how it works, server vs client architecture, tool discovery, and how Copilot uses MCP to connect to external systems.

Organization-level MCP policies, MCP registry URLs, controlling which MCP servers can be used, enterprise MCP governance.

Conceptual overview of the plugin system — what plugins are, how they bundle customizations, install/update lifecycle, and how they interact with other features.

How to write effective custom instructions, prompt files, and tips for guiding Copilot's responses to match your preferences and project standards.

Guidelines for responsible AI usage — limitations, security considerations, code review practices, and understanding AI-generated outputs.

The cloud-based async AI teammate — assign issues, spin up dev environments in GitHub Actions, create PRs automatically. The conceptual overview.

How to use coding agent — assign issues, track work in the Agents panel, review PRs, customize the dev environment, and manage agent tasks.

Create specialized agents for the cloud-based coding agent — shared across your organization via the .github-private repository.

Create agent skills for the cloud-based coding agent — same SKILL.md format, project-level skills in .github/skills/.

Connect MCP servers to the cloud-based coding agent — configure via repo settings, JSON config format for MCP server definitions.

Use hooks with the cloud-based coding agent — same .github/hooks/*.json format, runs in the Actions dev environment.

Organization-level policies for enabling/disabling Copilot features, coding agent, CLI, and managing member access.

Agent mode, chat panel, inline completions, next edit suggestions, multi-file edits, MCP servers, custom agents, subagents, vision — the flagship experience.

Agent mode, chat, completions, and code review in Visual Studio for .NET, C#, C++ development.

Agent mode (GA), chat, completions, custom agents, subagents, plan mode — IntelliJ, PyCharm, WebStorm, Rider, and more.

Agent mode GA for Eclipse — autonomous multi-step task execution for Java and JVM languages.

Agent mode GA for Xcode — autonomous coding for Swift and Objective-C with inline completions.

Inline completions and chat via the Copilot.vim plugin for terminal-native editors.

Use Copilot Chat directly on GitHub.com — ask about repos, PRs, issues, code, and GitHub features.

Copilot Chat on iOS and Android — ask questions, assign issues to coding agent on the go.

Compare Free, Pro, Business, and Enterprise plans — features, pricing, premium request quotas, and model access.

How premium requests work — model multipliers, what counts as a request, monthly quotas, and how different models affect your usage.

Individual plan billing details — Free tier limits, Pro plan features, premium request allowances, and overage handling.

Track premium request consumption, view usage dashboards, monitor team entitlements, and manage cost across your organization.

Curated collection of customization examples — custom instructions, agent configs, skill templates, and MCP server setups you can use as starting points.

Ready-to-use custom instruction templates — coding standards, framework-specific guidelines, testing patterns, and documentation requirements.

Step-by-step tutorial on creating hooks for predictable, policy-compliant execution — real-world examples for security scanning, logging, and guardrails.

Tutorial on using custom instructions to unlock the power of Copilot code review — configure review standards and enforce consistency.

Add custom instructions for all Copilot surfaces (not just CLI) — .github/copilot-instructions.md for repo-wide standards across VS Code, GitHub.com, and more.

The official Copilot CLI repository — README, changelog, issues, discussions, and the install script. File bug reports and feature requests here.

Official GitHub-maintained plugin marketplace — curated collection of plugins ready to install with /plugin marketplace add github/copilot-plugins.

Community-curated collection of Copilot resources — skills, plugins, custom agents, MCP servers, and examples. Also serves as a plugin marketplace.

The Agent Skills open standard specification — defines the SKILL.md format used by Copilot CLI, Claude Code, and other AI systems.

Anthropic's collection of agent skills — compatible with Copilot CLI's skills system. Ready-to-use skills for common tasks.

Third-party Copilot Chat extensions — Docker, Sentry, Jira, Confluence, and more. Build and deploy with the tools you love in Copilot Chat.

The main GitHub Copilot product page — features overview, demos, and getting started resources.

The AGENTS.md specification by OpenAI — defines the agent instructions format supported by Copilot CLI, Claude, and other AI coding assistants.

Comprehensive introduction to Copilot coding agent — what it is, how it works, assigning issues, customizing environments, and integration strategies.

Advanced strategies — offload tech debt, validate UI changes, experiment with branches, multi-agent workflows, and CI integration.

Changelog announcement — agent mode is now generally available for JetBrains IDEs, Eclipse, and Xcode, bringing autonomous coding to more developers.

Microsoft DevBlog — custom agents, subagents, plan mode, and new agentic capabilities for GitHub Copilot in JetBrains and Eclipse IDEs.

The complete GitHub Copilot blog — all announcements, tutorials, deep dives, and changelog entries in one place.