Assessment Guide

These assessments measure whether you can make sound decisions with an AI agent, not whether you can remember product trivia. Use them to establish a baseline, practice at each chapter boundary, and document growth.

Assessment path

1. Read the answer-free Term Sheet, Competency Framework, and Scoring Guide.
2. Before Chapter 1, complete and save the entry form in Paired Diagnostics.
3. After each chapter, complete and save its task in Chapter Checkpoints.
4. Use Finish and Save Before Review before opening worked patterns. Disclose any prior exposure instead of pretending it did not happen.
5. Self-score and revise. When possible, ask a peer to score the saved evidence without coaching.
6. After Chapter 11, complete the different, paired exit situations on the same competency scale.
7. Compare entry and exit results with the Competency Framework, then choose your next practice target.

The navigation route above does not include a direct answer-key item. It reduces accidental preview; it cannot and is not intended to prevent deliberate access.

The chapter tasks complement the Capstone. They are smaller competency checks; the capstone asks you to combine the competencies in one responsible build.

What is assessed

Code	Competency
C1	Explain AI-agent behavior and limits
C2	Frame and delegate bounded work
C3	Operate files, versions, and sessions safely
C4	Run the Builder Loop and verify results
C5	Design useful, maintainable context
C6	Select tools and manage trust boundaries
C7	Mechanize quality and orchestrate work
C8	Ship and operate a responsible, viable result

See the Competency Framework for observable behaviors at each level.

Fair-use conditions

• Use the lane that matches your current work. A no-code artifact and an engineer artifact can earn the same score when they provide equivalent evidence.
• Submit in writing, as an audio or video explanation, or through a live demonstration. Include links or timestamps so a reviewer can locate the evidence.
• Use screen-reader, captioning, speech-to-text, keyboard, display, or translation support as needed. The assessment does not reward typing speed or polished English.
• Work untimed unless a real constraint is part of the task. Take breaks.
• Use synthetic or public data. Never expose secrets, regulated data, client-confidential material, or personal data to complete an assessment.
• You may use an AI assistant except where a prompt asks for your independent judgment. Disclose what it contributed and verify the result yourself.
• If a service, account, payment method, or local installation is unavailable, use the documented simulation option. Score the quality of the decision and evidence, not tool access.

Keep an evidence folder

Create one folder in your course journal:

assessment/
  entry/
  chapter-01/
  chapter-02/
  ...
  chapter-11/
  exit/

Each folder should contain or link to the artifact, your score, the evidence for that score, feedback received, and one next action. Do not publish the folder unless every item is safe to share.