CCAO-F - Configuration and Knowledge Management (12% of the exam) - Section 5.1

Configure Claude Projects with instructions and knowledge sources.

Setting up a Project for repeated work: writing its instructions and adding the documents it should draw on, so every chat in the Project starts with the same context instead of re-pasting it.

Projectsproject instructionsproject knowledgeshared context across chats

Practice question for this objective

Free sampleConfiguration and Knowledge Managementmedium

A project manager at a software firm uses a shared Project to write fortnightly sprint updates. In one chat last week, she and Claude agreed that every update would open with a three-line risk summary. Today a colleague started a new chat in the same Project, and his draft has no risk summary. Which explanation best accounts for this?

  • AThe format was agreed inside one chat, and only instructions and knowledge are shared by every chat. Correct
  • BThe Project's knowledge files have grown too large for Claude to read the agreed format.
  • CThe colleague's chat is running a faster model that does not follow formatting agreements.
  • DClaude follows a formatting decision only when it is repeated in several earlier chats.
Decisions reached inside one chat do not become shared Project context; put lasting rules in the project instructions or knowledge so every chat applies them. A Project gives every chat the same starting point: its instructions and its knowledge. What two people agree in the back-and-forth of a single chat stays in that conversation, so a colleague opening a fresh chat starts without it. A rule the whole team should get every time belongs in the project instructions.

Why A is correct: Correct. Every chat in a Project starts from the same instructions and knowledge, but an agreement reached in the conversation of one chat is not part of that shared setup, so a colleague's new chat has no reason to apply it.

Why B is wrong: It is tempting to blame a crowded knowledge base when output drifts. But the risk summary format was never added to the knowledge, so the size of the files cannot explain why it is missing.

Why C is wrong: Model choice can affect depth of reasoning, so this sounds like a possible cause. Any model follows a clearly stated format; the problem is that the format was never stated anywhere the colleague's chat could see it.

Why D is wrong: This mistakes Claude for something that learns habits by repetition across chats. Repeating the decision in more chats would not make it shared context; writing it into the project instructions would.

See more CCAO-F practice questions, answers explained.

Exam traps in Configuration and Knowledge Management

Answers that look right on this material and are not. Each one is a distractor from a different question in the CCAO-F bank for this domain.

  • Keep the long conversation going and restate the corrected parking rule above each new query she pastes in.

    Why it is wrong: Restating the correction is tempting because it fixes the symptom she noticed. It leaves her in a conversation long enough to blur earlier detail, so other corrections and the reply style can drift in the same way.

  • Set up the Project as suggested, because a Project gives better answers than an ordinary chat.

    Why it is wrong: Projects are often presented as the more advanced option, so this feels like the careful choice. A Project does not make Claude reason better; its value is reusable context for recurring work, which this one-off task does not have.

  • Claude needs several days to learn a newly uploaded policy before it uses it

    Why it is wrong: It is tempting to think of uploads as training that takes time to sink in. Project knowledge is not training: it is reference material Claude can draw on straight away in the Project's chats, which is why her own test questions already return the new rules.

Examworthy is not affiliated with or endorsed by Anthropic. Original, blueprint-aligned practice material only.