When Claude Code is working, it narrates what it's doing using developer words. This is the glossary. Read it once. Recognize the words next time you see them and you'll know exactly what's happening.
These are the tools Claude Code has on its belt. When you see one of these words appear, that's Claude reaching for a specific capability — like a chef pulling a knife versus a whisk.
git status and tell me what's changed."These are the action words Claude uses as it narrates its work. When you see these, you know roughly what's happening under the hood.
/compact. Compresses the chat history into a dense summary so you can keep working without hitting the context ceiling.git diff. Shows the difference between two versions of a file, line by line.These come up constantly. Once you know what they mean, half of Claude Code's output makes sense.
/Users/brandon/CORTEX/notes.md is a path. Absolute = full address. Relative = "from where I am now."/compact or /clear./permissions. Tells Claude which tools it can use freely versus which require your approval per use.--- wrapped key-value pairs. Used in skills, agents, and commands..md. Uses # for headings, * for bullets, etc. CLAUDE.md, SKILL.md, and your CORTEX notes are all markdown.You don't need to write code, but these words will come up. Knowing what they mean lets you read what Claude says without guessing.
clientName = "Schneider" stores "Schneider" in a box labeled clientName."Hello" is a string. "Schneider Haus" is a string.["Allie", "Nate", "Avery"] is an array of three names.{name: "Brandon", role: "CVO"} is an object with two fields.The fastest reference. See it. Know what it means. Move on.