Whetstone.
Sign in
All courses
anthropicIntermediate

Claude Agents and Tools

Agents and Workflows is 14.7% of the CCDV-F exam and Tools and MCPs is another 10.6%, so a quarter of the paper lives in this course. It covers the four different Anthropic products that all claim to build agents and refuses to let you confuse them, the tool loop at wire level, tool_choice and the strict-tool-use placement trap, the built-in tool catalogue with its real prices and beta headers, Agent Skills including a version conflict the docs have not resolved, and MCP from the two-part connector requirement down to what a server actually exposes. It is honest where the documentation is ambiguous, because the exam is timed and knowing which facts are lookups is worth as much as knowing the facts.

10 lessons 210 min
CACrash course
Curriculum

10 lessons across 4 modules

Capstone

Spec one agent end to end, then run its loop by hand

The brief

You are building a research assistant. It reads a question, searches the web, runs a bit of code on the results, and writes a summary. Nothing exotic. The point is that every decision you make building it maps onto an exam objective.

Do it in two halves, and the first half is the one that matters.

Half one: the spec, on paper

Write a single markdown file that answers these, with a sentence of justification each. No code, no network, no account. This is the barn wifi half.

  1. Which of the four products runs this? Agent SDK, Client SDK, Tool Runner, or Managed Agents. Say what you gain and what you give up.
  2. Which model, and why? Name the context and output limits you are relying on, and say whether you need adaptive thinking.
  3. Which pattern is this? Reach for the vocabulary from Module 1. If it is a workflow, say so plainly rather than calling everything an agent.
  4. Which tools are built in and which are yours? For the built-ins, write down the cost. For yours, write the full tool definition including input_schema.
  5. Does strict go anywhere? If yes, say exactly which object it goes on. If a tool is coming from MCP, say what that changes.
  6. Where does it deploy? First-party API, Bedrock, Vertex, or Foundry. Then cross off every feature that platform does not have.

If step 6 kills a choice you made in step 1, that is the capstone working correctly.

Half two: the loop, by hand

Now run it with no SDK loop helper. Write the request body as JSON in a scratch file, then hand-simulate three turns:

  • Turn 1: your request, and the assistant response you expect back, with stop_reason set to whatever it should be.
  • Turn 2: the user message you send next, containing the tool result. Get the block type and the id field right from memory.
  • Turn 3: the final assistant turn, and the stop_reason that tells your loop to stop.

Then break it on purpose. Write the version where a tool throws, and show the exact shape you send back so the model can recover instead of your process crashing.

The bar

You are done when someone could read your file and implement the thing without asking you a single clarifying question, and when the failure turn is in there. The failure turn is the part everybody skips, and it is the part the exam asks about.

Your progress0%

0 of 10 lessons complete

anthropic
agents
tools
mcp
ccdv-f