Stage 2 · Lesson 1 of 7
How a request becomes a context window
When a user sends a message, the app builds the text the model will read: the (the developer's instructions), earlier conversation, retrieved documents, tool descriptions and tool results, then the user's message. This bundle is the context window.
The model reads all of it and produces the next . Ordinary code then takes the answer and shows it, stores it or acts on it. Security work starts by drawing this assembly for a real app: every source that adds text to the context is a way in, and every place the answer goes is a way out. Notice how many sources the user never sees, such as a web page fetched by a tool.
Common trapthinking the chat box is the only input, when emails, files and web pages the app reads on its own are often the bigger risk.
A question to check yourself, an AI practice session and 1 cheat sheet entry come with the full roadmap. Unlock
Resources for this stage
Start with· follow one of these
- CoursePaidGrokking Prompt Engineering for Software Engineers
Design Gurus · Beginner · Subscription
How system prompts, context and structured output really work, the ground every LLM risk stands on
- CourseFreeHugging Face Learn
Hugging Face · Intermediate
Free courses on how language models, embeddings and agents work, at an engineer's level