An support representative is configuring a Project for drafting customer-facing apology emails after service outages. The current custom instructions tell Claude to " be empathetic and casual, keep it short, and offer a goodwill credit. " The customer-service team has flagged that recent drafts have been too brief, have offered goodwill credits in cases where policy does not allow them, and have used contractions that the brand voice guide prohibits.
Which adjustment best aligns the configuration to the use case?

• A. Keep the existing instructions and add a sentence telling Claude to consult the brand voice guide and the goodwill-credit policy that are attached to the Project knowledge base.
• B. Move the brand voice guide and the goodwill-credit policy into the Project knowledge base and rely on the documents alone to govern tone, length, and credit decisions, without changing the custom instructions.
• C. Specify the brand voice attributes (empathetic, formal, no contractions), set a minimum length, and instruct Claude to recommend a goodwill credit only when the attached eligibility policy permits one.
• D. Replace " be empathetic and casual " with " match the brand voice guide, " remove the instruction to offer goodwill credits, and let each draft determine its own length based on the incident.

You are a knowledge manager structuring source material in the context window for a long Claude evaluation task.
Which two practices produce a more effective use of the context window? (Select two.) Each correct answer presents a complete solution.

• A. Mix sources together in a single block so the model can synthesize them freely.
• B. Place the most reference-critical material near the beginning or end of context.
• C. Mark each source clearly with delimiters so the model can attribute claims correctly.
• D. Repeat the same instruction at the start, middle, and end to reinforce it.
• E. Paste every available document into context regardless of relevance to the task.

Before sharing a Claude-generated brief, an analyst is applying the AI Fluency Framework Discernment competency to evaluate it.
Which review best reflects Discernment principles?

• A. Skip the source material check during review and rely on overall tone and logical flow as proxies for factual accuracy.
• B. Accept the brief on the basis that it reads fluently and the subject matter aligns with the analyst ' s professional knowledge of the topic.
• C. Conduct a systematic check of the brief against the task requirements, source material, and professional standards before sharing it.
• D. Review the opening paragraph and the conclusion of the brief, since these sections typically reflect the quality of the full document.

A knowledge worker is applying the AI Fluency Framework Description competency while revising an underperforming prompt.
Which revision best reflects Description principles?

• A. Specify the task scope, required output format, audience, and relevant constraints with sufficient clarity that the prompt produces a useful, reliable output on the next run.
• B. Remove the existing context from the prompt and submit only the bare task statement, eliminating the audience, format, and constraint information that Description principles call for in a clear specification.
• C. Add length to the prompt by inserting filler text that does not specify task scope, output format, audience, or constraints, leaving the original specification gaps unaddressed.
• D. Replace the specific instructions with generic phrasing that broadens the task scope, undoing the specification work that Description principles require for a reliable, useful output.