CCDV-F Agents and the Agent SDK Practice Question
An agent using the SDK encounters an error where it generates a tool call for a tool that does not exist in its configuration. Which adjustment to the agent's setup is most likely to resolve this hallucination?
⚠ Common exam trap
Candidates often try to fix hallucinations by adding more tools. This ignores the root cause: the model lacks clear guidance on which tools are actually available in the current context.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Refining the system prompt to clearly define the boundaries of available tools.
Hallucinating non-existent tools usually occurs when the model is aware of a capability but hasn't been given the structured definition for it, or when the system prompt is ambiguous. Refining the system prompt to explicitly list the available tools and their purposes helps ground the model's reasoning in the provided technical context.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Enabling 'force_tool_use' to ensure it only uses one of the existing tools.
Why it's wrong here
Forcing tool use might stop the model from providing a text response, but it doesn't necessarily stop it from hallucinating a tool name if it believes that tool is the 'best' fit for the task. The root cause is the model's misunderstanding of the available toolset, not just its decision to use a tool.
- ✓
Refining the system prompt to clearly define the boundaries of available tools.
Why this is correct
A well-defined system prompt serves as a 'source of truth' for the agent. By explicitly stating 'You only have access to tools X, Y, and Z,' and describing their specific use cases, you reduce the likelihood of the model attempting to invent a 'convenient' tool that doesn't actually exist in the SDK.
- ✗
Switching the model from Claude 3.5 Sonnet to Claude 3 Haiku.
Why it's wrong here
Smaller models like Haiku are generally more prone to hallucinations and have less sophisticated reasoning capabilities than larger models like Sonnet or Opus. Switching to a smaller model would likely exacerbate the problem of hallucinating tool names rather than solving it through improved grounding or better instruction following.
- ✗
Increasing the max_tokens parameter to allow for more reasoning.
Why it's wrong here
Increasing the token limit gives the model more space to talk, but it doesn't improve the accuracy of its tool selection. If the model is already hallucinating a tool, giving it more tokens might just result in a longer, more detailed explanation of why it is using that non-existent tool, without fixing the error.
Visual reference
About these practice questions
One of 257 original CCDV-F practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Anthropic exam blueprint
This CCDV-F practice question is part of Courseiva's free Anthropic certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the CCDV-F exam.