The short answer
Group the terms by purpose: LLM means language model; tokens are text units; context is available information; prompts are instructions; RAG adds retrieval; agents use tools; skills package instructions; MCP connects systems; a harness supports execution; APIs expose programmatic functions.
Try it in three steps
- Pick three unfamiliar terms from an AI product description.
- Write a plain explanation and one use for each.
- Keep general concepts separate from product names.
A practice prompt
Before you start
Compare terms in one small scenario instead of memorizing English first.
Practice material (fictional)
Task: find current service rules, then draft an answer. Identify model, retrieval, tool action, instructions, permissions and checks.Replace placeholders in the prompt with this material, or append the material if there are no placeholders. Reading time excludes doing the exercise.
Explain agent, skill and harness using a cooking analogy, then explain where the analogy breaks down.What to look for (not a recorded test result)
An LLM generates text; RAG includes retrieval; an agent can advance actions. Skills supply procedures, MCP standardizes connections and a harness supports execution and checks. They are not mutually exclusive products.
If it goes wrong
An analogy does not prove a product supports a feature; check its documentation.
Check your result
You can ask what a tool connects to, what it does and how its result is checked.
References and scope
The links below support the concepts or product navigation. The steps and exercises are our teaching suggestions, not official certification or measured customer outcomes. Product interfaces and availability may change. References checked on 2026-10-17.