BuildMidBackendAuto-graded

Build a prompt renderer that resists template injection

Supplied data must never turn into instructions, while missing values and optional sections still have to behave correctly.LLM prompts are assembled from a trusted template your team writes and variables that are frequently untrusted — user names, retrieved chunks, tool output, chat history. The template provides structure; the variables fill holes. Implement render(template, vars) with placeholder substitution, a conditional block and a loop — and make the substituted values inert data, so a variable containing template syntax cannot inject structure of its own.

35 min

python

Start lab

You're done when

The conditional block includes or omits its contents based on the named key.
The loop block repeats its body across a list.
Substituted values are inert — template syntax appearing inside a variable is not interpreted.
A missing variable is handled predictably rather than raising or leaking the raw placeholder.

What it measures

Prompt engineeringSeparating the trusted structure of a prompt from the untrusted content that fills it.

Context engineeringAssembling prompts from parts with conditionals and loops, so the structure stays legible as it grows.

Injection awarenessRecognising that a template engine re-interpreting substituted values is itself an injection surface.

Your environment

Terminal + editorA full shell in the browser

Claude CodePre-authenticated, no key setup

Isolated microVMYours alone, wiped afterwards

Start this lab

35 min

Start labYou'll sign in to Anthropos to start. The lab opens in a private cloud workspace with Claude Code ready to use.

AI Labs is part of Anthropos Workforce. Not on your plan yet? Book a demo