Yellow Belt · Promotion exam · PCE
Prompt & Context Engineer
Validates the ability to design how instructions, examples, documents, conversation history and tool results are arranged so that a model works consistently, and to evaluate and improve the results.
Daily practice · Free
10 questions · 20 min · No sign-up needed
Promotion exam · Proctored
30 questions · 45 min · Raw score of 70 or more · $3.99 / one attempt
Promotion exam detailsExam overview
- Questions
- 30
- Time
- 45 min
- Pass mark
- 70 pts
- Recommended experience
- White Belt (AI Literacy Foundation) fundamentals and at least 3 months of hands-on use of generative AI at work
Every question has five choices and is marked on an OMR answer sheet. Multiple-choice and true-or-false questions ask for the right combination from a box; ordering and matching questions ask for the right arrangement. Judgment questions (L3) are worth 3 points and all others 2; the raw score is converted to a 100-point scale and graded 1 to 9.
Exam domains
1. Instruction design25%
2. Examples and structure20%
3. Context engineering25%
4. Reasoning and multi-step tasks15%
5. Evaluation and improvement15%
Weights follow the Promotion exam. Practice and mock exams may mix domains differently.
What this exam verifies
- Design instructions that state the task goal, constraints and success criteria
- Make output more consistent with examples, structured tags and output formats
- Arrange long documents, conversations and tool results within a context budget
- Validate prompt changes with evaluation criteria and test sets
- Diagnose and fix failure patterns (missed instructions, broken formatting, overgeneralization)
Who it is for
- Planners, marketers and researchers responsible for the quality of AI output
- People who maintain internal prompt templates
- Practitioners at the stage before API development
Does not cover model training or building infrastructure.
Domain details
Domain 1. Instruction design
- Knowledge: Role, goal, constraints and success criteria, Positive instructions, Explaining the reason, Roles of system, developer and user messages
- Skills: Turn vague requirements into verifiable instructions, Resolve conflicting instructions
Domain 2. Examples and structure
- Knowledge: Diversity of few-shot examples, Structured tags (XML, etc.), Output schemas, Prefill and format locking
- Skills: Reduce bias introduced by examples, Specify machine-readable output formats
Domain 3. Context engineering
- Knowledge: Context window budget, Ordering of long documents, Asking for quotes first, Conversation summarization and compaction, How retrieval augmentation works
- Skills: Include only the information that is needed, Have the model quote evidence from long documents
Domain 4. Reasoning and multi-step tasks
- Knowledge: Reasoning models and reasoning effort, Task decomposition, Prompt chaining, Self-review
- Skills: Split multi-step tasks into step-by-step prompts, Match reasoning effort to cost and latency
Domain 5. Evaluation and improvement
- Knowledge: Evaluation criteria (rubrics), Test sets, Regression, Limits of LLM evaluators
- Skills: Compare before and after on the same set, Classify failure cases to decide the next fix
Official documentation
Anthropic
Yellow Belt Course · Prompt & Context Engineer
Practice, mock exams and every course book are free. Pay only when you take an official test.