Skip to content

Resources · Glossary

Glossary

Every term used on this site, in plain English. If a page uses a word that isn't here, tell me and I'll add it.

Agent — an AI that can do things (read files, write drafts, run steps), not just chat.

Automation — a job set up to run without you starting it each time.

Commit — in Git, a saved snapshot of your project with a note about what changed.

Context / context window — everything the AI can see right now: your message, the conversation so far, attached files. If it's not in the window, the AI doesn't know it.

Folder agent — an agent that works inside one folder on your computer (Claude Code, Cursor).

Git — a tool that tracks versions of a project folder so you can save, share, and undo.

Hallucination — when AI confidently states something false. The fix is sources you can check, not trust.

LLM (large language model) — the technology behind AI chat tools. You can just say "the AI."

MCP — the standard plug that connects AI tools to apps like your calendar or files. Think "USB for AI."

Model — a specific AI (Claude, GPT, Gemini are families of models). Newer models are generally more capable.

Playbook — on this site: one useful job packaged so your AI can install it, remember it, and run it by interview. See how playbooks work.

Prompt — your request to the AI. Nothing fancier than that.

Project instructions — rules you write once that apply to every chat in a project.

Skill — the file a playbook installs into your vault so the AI can run the job later.

Subagent — a helper AI that the main AI hands part of a job to.

Terminal — the text box where you type commands to your computer.

Token — the small chunks of text AI reads and writes; pricing and limits are counted in them. You rarely need to think about tokens directly.

Tool — anything the AI can use besides talking: file access, web search, an app connection.

Vault — a folder of plain-text notes that serves as your (and your AI's) long-term memory.

Workflow — the full path a job takes: trigger → input → AI step → check → approval → delivery.

Next step: Learning links