Microsoft Developing in Agentic AI Systems (GH-600) Exam Questions

Microsoft Developing in Agentic AI Systems Exam Objectives

1. 1

   1.0 Prepare agent architecture and SDLC processes (15-20%)

   • 1.1Integrate agents into the software development lifecycle (SDLC)
   • 1.2Identify steps for agents to perform
   • 1.3Identify and mitigate common anti-patterns in agents
   • 1.4Define inputs, outputs, and success criteria for agents
   • 1.5Define boundaries between planning, reasoning, and action
   • 1.6Configure agent planning to be distinct from agent execution
   • 1.7Configure an agent to output a structured plan
   • 1.8Validate agent plans
   • 1.9Prevent agent action until the agent checked and approved
   • 1.10Configure observability and control for autonomous agents
   • 1.11Plan and implement the degree of agent autonomy, including guardrails
   • 1.12Configure agent to produce inspectable artifacts within standard development tooling
   • 1.13Configure human intervention for autonomous agents without slowing delivery
2. 2

   2.0 Implement tool use and environment interaction (20-25%)

   • 2.1Select and configure agent tools
   • 2.2Identify required tools
   • 2.3Configure agent tools
   • 2.4Configure agent tool permissions
   • 2.5Configure MCP servers
   • 2.6Add an MCP server as a tool to an agent
   • 2.7Configure a GitHub remote MCP server
   • 2.8Configure the MCP registries
   • 2.9Configure MCP allow lists
   • 2.10Integrate agents within development environments
   • 2.11Evaluate the execution context for an agent
   • 2.12Configure an agent's scope to a specific repository
   • 2.13Configure an agent to be invoked in a CI workflow
   • 2.14Configure an agent to use branch-based scope
   • 2.15Enable an agent to perform autonomous actions, including creating branches and pull requests
   • 2.16Configure an agent to handle environment-specific constraints
   • 2.17Operate agents with safe execution paths and robust error handling
   • 2.18Implement error handling
   • 2.19Implement retries
   • 2.20Implement rollbacks
   • 2.21Implement escalation paths
   • 2.22Implement traceability and accountability for agent actions
3. 3

   3.0 Manage memory, state, and execution (10-15%)

   • Task 3.1: -relevant information Define memory expiration, pruning, and reset rules Persist agent state and manage context drift Capture
   • Task 3.2: progress and decisions as durable artifacts Resume agent work without repeating steps or diverging from prior decisions Detect and correct drift during extended agent execution Ensure continuity of agent memory and state across tools and environments Share agent state Prevent conflicting context Prevent stale context
4. 4

   4.0 Perform evaluation, error analysis, and tuning (15-20%)

   • 4.1Define success criteria and evaluation signals for agent tasks
   • 4.2Specify expected outcomes and operational constraints for agent tasks
   • 4.3Identify qualitative and quantitative evaluation signals to evaluate agents
   • 4.4Align evaluation criteria with development intent
   • 4.5Generate evaluation signals by using automated scanning tools
   • 4.6Analyze agent failures and identify root causes
   • 4.7Identify failures by using logs, plans, traces, outputs, and workflow artifacts
   • 4.8Classify root causes, including reasoning errors, tool misuse, and context or environment issues
   • 4.9Tune agent behavior based on evaluation results
   • 4.10Revise instructions, workflows, or constraints
   • 4.11Refine memory usage
   • 4.12Refine tool usage and tool access
5. 5

   5.0 Orchestrate multi-agent coordination (15-20%)

   • 5.1Operate and manage multi-agent workflows
   • 5.2Apply an orchestration pattern to coordinate multiple agents
   • 5.3Configure agent isolation for parallel execution
   • 5.4Detect and resolve agent conflicts, including overlapping code changes, duplicated effort, and contradictory outputs
   • 5.5Configure observability for multi-agent behavior by using logs, artifacts, and operational signals
   • 5.6Configure multi-agent workflows to produce artifacts suitable for review and audit
   • 5.7Document key decisions, handoffs, and outcomes across agents
   • 5.8Perform post-hoc analysis of multi-agent behavior
   • 5.9Detect and respond to multi-agent failures and degraded behavior
   • 5.10Identify failed, partial, or stalled agent executions
   • 5.11Respond to degraded behavior or coordination across agents
   • 5.12Implement multi-agent recovery patterns, including rollback and human-in-the-loop
   • 5.13Manage the lifecycle of agents within multi-agent workflows
   • 5.14Add agents to existing multi-agent workflows
   • 5.15Update, reconfigure, or replace agents without disrupting active workflows
   • 5.16Retire agents while preserving auditability and workflow continuity
6. 6

   6.0 Implement guardrails and accountability (10-15%)

   • 6.1Define autonomy levels
   • 6.2Classify agent actions by operational, security, and compliance risk to right-size human interventions
   • 6.3Assign autonomy levels to maximize delivery speed while remaining compliant with organizational security and Responsible AI standards
   • 6.4Implement guardrails and human-in-the-loop workflows
   • 6.5Identify the subset of actions that require human judgment
   • 6.6Block actions that violate defined security, compliance, or Responsible AI policies
   • 6.7Scope permissions and execution contexts to enforce least-privilege access
   • 6.8Require explicit authorization or controlled paths for irreversible or compliance-sensitive changes
   • 6.9Preserve execution velocity by minimizing approvals that do not materially reduce risk