Module 1: The Terminal-First Social Workflow

Here’s what social media management usually looks like. You finish something interesting. You open Twitter, write something half-formed. Then LinkedIn, write something different and longer. Then remember you wanted to post to Threads as well. Three tabs, three contexts, three rewrites of roughly the same idea — none of them quite right because you’re thinking about the platform instead of the thought.

That’s the dashboard problem. Each platform wants your attention inside it, which means you’re always in reaction mode. You’re adapting yourself to the interface instead of the interface adapting to your work.

The terminal-first approach flips this. You stay in your working environment. You tell Claude what you built or figured out. It drafts per-platform versions. You approve them. They queue to Buffer. Done.

Why the Terminal Wins

It’s not about being a developer. It’s about where your attention lives when you’re doing actual work.

If you’re already in a terminal building something, switching to a browser tab breaks the context. You lose the thread of what you were doing, spend a few minutes trying to summarise it for LinkedIn, produce something mediocre, and close the tab feeling vaguely dissatisfied.

If posting is a command — something you run in the same environment where the work happened — you don’t break context. The work and the post about the work are in the same place. The friction drops enough that you actually do it.

The Post-to-Buffer Pattern

The core workflow is simple:

1. You do something worth sharing — a problem solved, a thing built, a lesson learned
2. You run the social-media agent or invoke the post-to-buffer skill
3. Claude drafts versions for each platform, formatted appropriately for each one
4. You review, edit if needed, approve
5. The posts queue to Buffer via the GraphQL API
6. Buffer publishes them at your scheduled times

The skill handles the platform differences. LinkedIn wants professional framing and longer context. Threads wants something that reads naturally in a feed — shorter, more direct. X has the character limit. You don’t think about any of that. You just describe what you did.

Claude Code as Your Social Media Assistant

The reason this works is that Claude Code has access to your context in a way no social media tool does.

If you just ran a build, Claude can read the output. If you solved a problem, it can read the code you wrote. If you wrote a blog post, it can read that too. The content you’d spend ten minutes summarising for a LinkedIn post is already right there — Claude just needs to extract the shareable angle.

This is content extraction, not content creation. You’re not inventing things to say. You’re making visible the things you’re already doing.

What the Agent Actually Does

The social-media agent in this system runs with Read, Write, Bash, and the post-to-buffer skill in its toolset. When you invoke it, it:

• Asks what you want to post about (or reads recent work from context)
• Drafts per-platform versions with appropriate length and tone
• Shows you the drafts for approval
• Calls post-to-buffer to queue the approved posts via Buffer’s GraphQL API

You never open Buffer’s interface. You never switch tabs. You stay in the terminal.

The Mental Shift

The biggest change isn’t technical. It’s stopping thinking of social media as a separate activity that requires a separate tool and a separate frame of mind.

When posting is integrated into your workflow — same environment, same tools, same session — it stops feeling like content marketing and starts feeling like documentation. You did a thing. You recorded what you did. The recording happens to be in a format other people can read.

That’s the frame. Everything in the rest of this workshop builds on it.

---

The next module covers Buffer’s GraphQL API and how the post-to-buffer skill actually works.