Fabric, fzf, tmux, SSH and pass: CLI workflow tutorial

This tutorial covers five command-line tools, with Fabric providing the AI workflow. The speaker demonstrates calling an LLM, applying reusable prompts called patterns, and chaining requests through shell pipes. One example retrieves a YouTube transcript, extracts ideas, creates a summary and writes a micro essay. Streaming displays the generated output as it arrives.

The speaker says his Fabric configuration uses Anthropic Opus 4.7 through the Anthropic API. This demonstration therefore sends content to a remote service; it does not establish how to run models locally or use Fabric offline. He also suggests reading the pattern files to study prompt construction.

The surrounding terminal examples cover fzf file search and command history, plus tmux sessions, windows and split panes. The speaker recommends learning history, find and grep before relying on fuzzy search, since remote machines may lack it. He favors standard tmux bindings for working across machines.

The SSH example connects a Fedora Linux machine to an Ubuntu virtual machine and forwards local port 8080 to a web server in a remote Podman container. Access stops when the tunnel closes. The pass demonstration generates passwords, stores them as local GPG-encrypted files and pushes those files to Git. The speaker's productivity and security endorsements are personal claims, rather than measured comparisons.