$ man context-wiki/github-repos

GitHub stores your Git repo in the cloud. That is the core function. Your code is not backed up until you push to GitHub. Your team cannot see your work until you push. Automated deployments do not trigger until you push. GitHub is the warehouse where your sealed boxes (commits) live. You can pick up your work from any machine. Clone the repo, and you have the entire history, every file, every commit. If my laptop dies tomorrow, I clone the repo on a new machine and I am back in 5 minutes. That is not paranoia. That is infrastructure.

PATTERN

Public repos are visible to everyone on the internet. Anyone can read your code, clone it, and learn from it. Good for open source projects, portfolios, and educational content. Private repos are locked down. Only people you invite can see them. Good for client work, proprietary code, and anything with sensitive data. My GTM OS repo is on GitHub. Every skill, every voice file, every partner workflow. It is the single source of truth. I keep it private because it contains partner-specific research and campaign strategies. But the website code deploys from it publicly. The repo is private. The deployed sites are public. GitHub and Vercel handle that separation automatically.

PRO TIP

.gitignore is a file that tells Git what to ignore. Add .env to your .gitignore and Git will never track your environment variables. Add node_modules/ and Git skips your dependency folder (which can be thousands of files). This is critical for security. API keys, database passwords, MCP tokens. These cannot live in your repo. If your .env file gets committed to a public repo, those keys are exposed to the entire internet. Bots scan GitHub constantly for leaked credentials. The rule: if it is a secret, it goes in .env. If it is in .env, it goes in .gitignore. No exceptions.

PATTERN

GitHub is not just storage. It is the backbone of your context engineering system. Your context repository lives on GitHub. Your CI/CD pipeline triggers from GitHub pushes. Your deployment previews generate from GitHub branches. Your commit history documents what you built and when. When I run the daily tracker, it scans git commits to count what I shipped. When I run /deploy, it pushes to GitHub and Vercel auto-deploys. When I onboard a new partner, the research goes into a GitHub folder that every future agent session can access. GitHub is the infrastructure that makes context persistent across sessions, machines, and agents.