Pricing
AI Platform Engineering
Phase 6Week 156 machine-verified checks · 1 tutorial pack

Prompting as Craft

Interface design, not a career.

Why this phase exists

You need enough prompting to define interfaces well: system design, schemas, tool descriptions. You do not need to be a prompt influencer. This phase gives you the craft and then explicitly tells you to stop studying it and go build.

What you learn

  • System design in prompts: roles, constraints, refusal behaviour
  • Structured outputs and schema-first tool definitions
  • Untrusted-input envelopes: never let retrieved text act as instructions
  • Prompt injection: a worked attack and the defence that stops it
  • What NOT to study: infinite prompt courses and framework fashion
Outcome: Your prompts are specs: reviewable, testable, versioned.

Tools you will actually touch

JSON SchemaTool/function callingInjection test suitesPrompt versioning

The assignment pack

Scenario

Your RAG assistant reads customer-supplied documents. One of those documents contains instructions telling the assistant to reveal system configuration. Harden the prompt layer so retrieved content can never act as instructions, and prove it with a test suite.

Definition of done
  • All tool definitions are schema-first with validated structured outputs
  • Untrusted content is delimited and provenance-tagged before it reaches the model
  • An injection test suite with at least ten attack variants, all failing to breach
  • Prompts are versioned and diffable, with a review process documented
How it is graded
Fail

Prompts are strings in code. Retrieved text is concatenated in directly.

Pass

Schema-first tools, untrusted-input envelopes, passing injection suite, versioned prompts.

Strong pass

The attack suite includes novel variants the student authored, and refusal behaviour is tested as carefully as success behaviour.

The tutorial pack

Break your assistant, then fix it

  1. 1Write a tool definition as a JSON Schema and validate the model output against it
  2. 2Plant an instruction inside a retrieved document and watch the assistant obey it
  3. 3Add delimiting and provenance tagging for untrusted content
  4. 4Re-run the attack and confirm it fails
  5. 5Build the injection suite and wire it into CI
The failure you will hit on purpose

Successfully prompt-inject your own assistant through a retrieved document. Most engineers do not really believe this works until they do it to themselves.

What you measure at the end
  • Injection suite pass rate (target: all attacks blocked)
  • Structured output validation failure rate under adversarial input
  • Refusal behaviour correctness on out-of-policy requests

You must be able to answer

AI use is mandatory on every assignment. The integrity mechanism is defend-your-work, so these are the questions that decide whether the work counts.

  • Show the attack that worked before your fix. Why did the fix stop it?
  • What is the difference between your system prompt and your policy layer, and which do you trust?
  • How would you review a prompt change in a PR?

Want to be walked through this phase?

The packs are delivered with live teaching, reviewed assignments, and a defence.

Train with me