Platform engineering has become a critical discipline for enterprises drowning in toolchain sprawl. Instead of building separate portals that developers ignore, the smartest teams are leveraging tools developers already use—like GitHub. This Q&A explores how golden paths, IssueOps, and Copilot integration create an invisible, high-velocity platform that slashes onboarding time and eliminates manual ticketing.
1. What is the biggest problem platform engineering solves?
Toolchain sprawl is the silent productivity killer. Developers at typical enterprises waste up to 40% of their time wrestling with infrastructure instead of writing code. They juggle outdated Backstage portals, slow ServiceNow tickets, and stale documentation. Platform engineering addresses this by offering a coherent, internal developer platform that automates repetitive tasks, enforces best practices, and provides self-service capabilities. However, many teams fail by building yet another portal that requires separate logins and maintenance. The real solution is to meet developers where they already are—in their everyday tools like GitHub. A well-designed platform is invisible; it removes friction without adding cognitive load.
2. Why should GitHub be the core of your IDP?
Most platform teams think they need a dedicated portal—backstage instances, custom dashboards—but developers already spend 8+ hours daily inside GitHub. By layering platform capabilities onto GitHub, you eliminate the "another tab, another login" problem. GitHub provides built-in CI/CD via Actions, code review workflows, issue tracking, and a rich ecosystem of integrations. Instead of building connectors between your IDP and GitHub, simply use GitHub as the single pane of glass. Patterns like golden-path repos and IssueOps transform native GitHub features into a full-featured platform. This approach reduces maintenance overhead, speeds up adoption, and lets your team focus on creating opinionated starter templates and automation rather than custom UI.
3. What are golden-path repos and how do they work?
Golden-path repos are opinionated starter templates that give developers a fast, recommended path without locking them in. They include pre-configured CI/CD pipelines, linting, security policies, and—crucially—Copilot context files. For example, a copilot-instructions-starter repo includes a .github/copilot-instructions.md that teaches Copilot your org’s coding conventions. Another example is copilot-agent-starter, which scaffolds custom Copilot CLI agents with extension architecture. When a new developer creates a repo from your golden path, they inherit the right workflows and settings instantly. This cuts onboarding from days to minutes. Golden paths are not rigid; like Rails conventions, they provide a default so good that deviations are rare. They accelerate teams while preserving freedom.
4. How does IssueOps replace traditional ticketing systems?
IssueOps eliminates the slow, multi-day journey through ServiceNow by turning GitHub Issues into the self-service interface for platform operations. Need a new environment? Open an issue with a specific label like provision:staging. Need a database? Use a structured issue template that asks for database name, region, and size. A GitHub Actions workflow picks up the issue, runs automation (e.g., Terraform or custom scripts), and posts the result as a comment. This approach keeps everything inside the developer’s existing workflow—no separate portal, no ticket transfer, no waiting. It also creates an auditable trail of changes. The repository gh-aw-overview demonstrates this pattern clearly. IssueOps reduces provisioning time from days to minutes and aligns with GitOps principles.
5. How can Copilot be integrated into a GitHub-based platform?
Copilot integration is a game-changer for platform engineering. By embedding context files and custom agents into your golden-path repos, you ensure every developer gets AI assistance aligned with your organization’s standards. The copilot-instructions-starter template sets up .github/copilot-instructions.md so Copilot understands your codebase’s preferred libraries, naming conventions, and architecture. The copilot-agent-starter lets you create custom CLI agents that can scaffold projects, run tests, or enforce policies. For advanced use, copilot-life-os-starters provide kits for building agentic systems on top of Copilot CLI. This means Copilot doesn’t just autocomplete code—it becomes a guided assistant that helps developers follow golden paths naturally. No separate plugin or portal needed; it’s all within the IDE or terminal.
6. What does a typical implementation look like?
A practical implementation starts by identifying the most painful developer workflows—like onboarding, environment provisioning, or dependency upgrades. Then, create a golden-path repo for each common service (e.g., microservice, frontend, data pipeline). Each repo includes .github/copilot-instructions.md, a copilot-agent.yml for custom agents, starter CI/CD workflows, and issue templates for common operations. Next, set up IssueOps: define issue templates with type: provision or type: decommission labels, and create Actions workflows that trigger on issue creation. For example, when a developer opens a new-environment issue, an Action runs Terraform to create the environment and posts the endpoint back. The result is a fully self-service platform living entirely inside GitHub. Subscribe to the htek.dev newsletter for step-by-step configs and full code examples.
7. How does this approach streamline developer onboarding?
Onboarding traditionally involves provisioning accounts, cloning repos, configuring IDs, and reading pages of documentation. With a GitHub-based IDP, a new developer simply gets access to the team’s GitHub organization. They pick a golden-path repo (e.g., microservice-starter), click “Use this template,” and automatically inherit all the necessary CI/CD pipelines, Copilot context, linting rules, and security policies. Copilot, already configured via copilot-instructions.md, starts offering relevant suggestions from day one. No need to manually set up environment files or request database access—IssueOps handles that. The onboarding time drops from multiple days to just 10–20 minutes. The developer stays in their familiar GitHub and IDE environment, feeling productive immediately. This reduces dependency on platform teams for routine tasks and lets teams scale faster.