GitLab Certified Duo Agent Platform Associate

Last checked November 10, 2026 against the official GitLab exam page.

1. Question 1Using DAP: Agentic Chat and MCP

   A product manager asks Duo to read a discussion thread, open a follow-up issue with a checklist and apply the right labels. Which capability is designed to carry out a request like this?

   1. ACode completion, which predicts the next lines while a developer types
   2. BThe GitLab Credits dashboard, which tracks monthly usage of features
   3. CAgentic chat, which plans multi-step work and uses GitLab tools to act
   4. DThe AI Catalog, which lists agents and flows so teams can discover them
   Show answer and explanationHide answer

   Correct answer: C

   Agentic chat can plan a multi-step request and use GitLab tools to create and update objects such as issues and labels. Code completion is an editor feature, the credits dashboard reports usage, and the AI Catalog is for finding agents and flows rather than running a task.

2. Question 2Building Custom Agents and Flows

   A team's custom agent reviews code for accessibility problems, but its system prompt only says 'Be helpful.' Which change most improves how consistently the agent behaves?

   1. ADefine its role, the scope of the review, the expected output format and what it must not do
   2. BRemove all instructions so the model can decide its own role and scope on each run
   3. CPaste every project document into the prompt so the agent knows everything about the project
   4. DKeep the prompt to a single greeting so the responses stay conversational
   Show answer and explanationHide answer

   Correct answer: A

   A system prompt sets the agent's role, scope, output format and boundaries, and those specifics make its behaviour consistent across runs. Removing instructions, pasting every document in, or reducing the prompt to a greeting each leave the agent without the guidance it needs.

3. Question 3DAP Fundamentals

   A team wants one automated process in which a triage step runs first, a planning step runs second and a merge request step runs last, with each step handled by a different agent. Which layer should they configure to sequence these steps?

   1. AAn MCP server, which exposes tools that any connected client may call
   2. BA foundational agent, which handles a single role using a fixed system prompt
   3. CThe AI Catalog, which lists agent definitions so others can find them
   4. DThe flow, which chains several agent steps into one ordered process
   Show answer and explanationHide answer

   Correct answer: D

   Flows are the layer that sequences several agent steps into one ordered process. A foundational agent covers one role on its own, the AI Catalog only lists and shares definitions, and an MCP server exposes tools without deciding the order of work.

4. Question 4Code Creation, Review, Security and Best Practices

   A developer types a comment that describes the function they want, and the IDE then drafts the full function below it. Which Code Suggestions mode is at work?

   1. ACode completion, which predicts the next few tokens of code already being typed
   2. BCode generation, which produces new code from a natural-language description
   3. CSecurity scanning, which flags insecure patterns in code that was just written
   4. DMerge request review, which comments on the changes after a commit is pushed
   Show answer and explanationHide answer

   Correct answer: B

   Code generation takes a natural-language description, such as a comment, and produces new code from it. Code completion predicts continuations of code the developer is already writing, while security scanning and merge request review are separate features that analyze code rather than write it.

5. Question 5Using DAP: Agentic Chat and MCP

   A security lead wants help triaging vulnerability findings from the latest pipeline and suggested remediation steps, inside a chat session. Which option is the best fit?

   1. AThe GitLab Credits report, which lists the cost of each pipeline run in the group
   2. BA custom agent created with an empty tool list and no access to project context or findings
   3. CCode completion, which suggests lines of code while a developer types in the editor window
   4. DThe Security Analyst foundational agent, built to triage findings and suggest fixes
   Show answer and explanationHide answer

   Correct answer: D

   The Security Analyst is a foundational agent designed for security work such as triaging findings and suggesting fixes. A custom agent with no tools cannot see the findings, code completion works inside the editor while typing, and a credits report is about cost rather than analysis.