You are implementing a custom tool for your Claude agent. The tool needs to interact with an external pricing service that returns product data.
Which of the following best practices would you apply as you develop this tool?

• A. Omit the tool description and let the model infer when to use the tool based on the tool's name and the rest of the prompt context.
• B. Implement the tool with no error handling and let the agent loop catch failures whenever the pricing service returns an error during operation.
• C. Define the tool with a clear schema, write a precise description for when to call it, and handle pricing service errors explicitly.
• D. Define the tool with a loose schema and let the model interpret the inputs flexibly on each call the agent makes.

A teammate has asked you to explain when a Skill would be the right choice over an MCP server. The teammate is unsure how the two differ in practice when both can be reused across teams.
How would you explain the distinction?

• A. A Skill and an MCP server are equivalent extension mechanisms that the team can use interchangeably for any reusable capability that needs to be accessible across teams.
• B. A Skill is the older mechanism and an MCP server is the newer one, so the team should prefer an MCP server for any reusable capability that the team builds going forward.
• C. A Skill is preferable for cross-team reuse because it loads more efficiently than an MCP server during normal operation in the team's typical multi-team workloads.
• D. A Skill bundles prompts, scripts, and data into a package the model loads as a unit while an MCP server exposes resources, tools, and prompts through a standard client interface.

Your Claude application has been running for several conversation turns, and you notice the model occasionally references information that was discussed many turns ago but is no longer relevant. You suspect context drift is causing the model to weight stale content too heavily.
How would you address the drift?

• A. Increase the context window size so all turns of the conversation remain visible to the model in full detail.
• B. Truncate the conversation so the model sees only the most recent turn during each subsequent response.
• C. Apply compaction to summarize older portions of the conversation so the gist remains while the specifics carry less weight.
• D. Reset the conversation after every turn so the model loses all prior turns when generating a response.

You are deciding between Claude models for a task. The team has identified three relevant tradeoff dimensions: quality, latency, and cost.
The right model is the one that...

• A. Satisfies the task's latency requirement first, then is evaluated against quality and cost thresholds to confirm the selection is acceptable across all three dimensions.
• B. Meets the task's quality requirements at an acceptable latency, with cost reviewed separately once the quality and latency bar has been established.
• C. Fits the task's quality, latency, and cost requirements together, recognizing that improving one dimension typically affects the others.
• D. Meets the task's cost target within a defined latency budget, with quality validated against a representative sample of inputs after the model is selected.

A teammate has asked you to explain why the team's Claude application is billed for output tokens at a different rate than input tokens. They had assumed the rate was the same for both.
How would you explain the difference?

• A. Output tokens are typically billed at the same rate as input tokens, and the apparent rate difference is a billing error to report to Anthropic.
• B. Output tokens are typically billed at a higher rate than input tokens, and cost models for the application should reflect both rates separately.
• C. Output tokens are typically billed at a lower rate than input tokens, because output tokens are cheaper to produce than input tokens are to process.
• D. Output tokens are not billed at all, because cost is determined entirely by the input tokens sent to the model on each request.