What is the loop continue condition?

stopreason == "tooluse"; exit on any other value.

What is the current stopreason values?

endturn, maxtokens, stopsequence, tooluse, pauseturn, refusal, modelcontextwindowexceeded.

What is the pauseturn?

Server-tool loop hit its limit (default 10 iterations); send the response back as-is.

Returned as a normal HTTP 200; stopdetails names the category.

What is the parallel results?

One toolresult per tooluse, all in one user message, before any text.

What is the tool failure?

Return toolresult with iserror: true; the loop keeps running.

maxturns (tool-use turns) and maxbudgetusd; no limit by default.

What is the result field?

Present only on subtype success.

Agentic Architecture & Orchestration27% of the exam

Agentic loops and stop_reason control

How a client-tool agentic loop works on the Messages API and in the Agent SDK: continue on stop_reason tool_use, branch on every other value, append full assistant turns and all tool_result blocks, and keep turn and budget caps as backstops that report the run as incomplete.

Official exam objective

Design and implement agentic loops for autonomous task execution

~8 min read · 11 practice questions · updated Oct 22, 2026

What the exam expects

Write loop control that continues on stop_reason "tool_use" and handles every other value it exits on.
Append each assistant response and its tool_result blocks to one running messages list.
Answer parallel tool calls correctly: one tool_result per tool_use block, all in the next user message, nothing after them.
Reject the named anti-patterns: parsing natural-language phrases, treating any text block as completion, or using an iteration cap as the primary stop.
Choose model-driven tool selection over hard-coded sequences for open-ended work, and read Agent SDK result subtypes.

Key terms: agentic loop (send, inspect stop_reason, run tools, return results, repeat); stop_reason (the response field that says why Claude stopped generating); tool_result (the user-turn block that answers a tool_use block by tool_use_id); ResultMessage (the Agent SDK message that ends the loop, with a subtype); backstop (a turn or budget cap that only catches runaway runs).

Key facts

Fact	Detail	Source
Loop continue condition	`stop_reason == "tool_use"`; exit on any other value	How tool use works (opens in a new tab)
Current `stop_reason` values	`end_turn`, `max_tokens`, `stop_sequence`, `tool_use`, `pause_turn`, `refusal`, `model_context_window_exceeded`	Stop reasons and fallback (opens in a new tab)
`pause_turn`	Server-tool loop hit its limit (default 10 iterations); send the response back as-is	Stop reasons and fallback (opens in a new tab)
`refusal`	Returned as a normal HTTP 200; `stop_details` names the category	Stop reasons and fallback (opens in a new tab)
Parallel results	One `tool_result` per `tool_use`, all in one user message, before any text	Parallel tool use (opens in a new tab)
Tool failure	Return `tool_result` with `is_error: true`; the loop keeps running	Tutorial: Build a tool-using agent (opens in a new tab)
SDK limits	`max_turns` (tool-use turns) and `max_budget_usd`; no limit by default	How the agent loop works (opens in a new tab)
`result` field	Present only on subtype `success`	How the agent loop works (opens in a new tab)

Loop continue condition

Detail: stop_reason == "tool_use"; exit on any other value
Source: How tool use works (opens in a new tab)

Current stop_reason values

Detail: end_turn, max_tokens, stop_sequence, tool_use, pause_turn, refusal, model_context_window_exceeded
Source: Stop reasons and fallback (opens in a new tab)

pause_turn

Detail: Server-tool loop hit its limit (default 10 iterations); send the response back as-is
Source: Stop reasons and fallback (opens in a new tab)

refusal

Detail: Returned as a normal HTTP 200; stop_details names the category
Source: Stop reasons and fallback (opens in a new tab)

Parallel results

Detail: One tool_result per tool_use, all in one user message, before any text
Source: Parallel tool use (opens in a new tab)

Tool failure

Detail: Return tool_result with is_error: true; the loop keeps running
Source: Tutorial: Build a tool-using agent (opens in a new tab)

SDK limits

Detail: max_turns (tool-use turns) and max_budget_usd; no limit by default
Source: How the agent loop works (opens in a new tab)

result field

Detail: Present only on subtype success
Source: How the agent loop works (opens in a new tab)

Sources for this section (5)

Check yourself

Why must the loop check stop_reason rather than whether the reply contains text?
When the response has stop_reason: "pause_turn", what should the application send next?
How must the loop answer a response with three tool_use blocks?
Why should code check the Agent SDK ResultMessage subtype before reading result?
When does a model-driven agent fit better than a hard-coded tool sequence?

The questions are free to read. Answers with their sources and self-grading are included with the pack.

Keep reading with the Claude Certified Architect - Foundations pack

Core concepts
Exam traps
Worked scenario

The pack includes:

All 30 lessons of the Claude Certified Architect - Foundations guide, with reading progress
Interactive decision walks and step-by-step sequence steppers
Inline practice questions and one-click topic practice (11 questions for this lesson)
An official source cited for every claim
An AI tutor that answers from the guide
Unlimited timed practice exams

Get the pack Already own it? Sign in

Practice for the Claude Certified Architect - Foundations exam

Certifyforce is an independent study resource and is not affiliated with, endorsed by or sponsored by Anthropic. Claude is a trademark of Anthropic.