Search...

Quick Feedback

Agentic AI Developer (GH-600) Study Guide

Domain-by-domain study guide for GitHub Certified: Agentic AI Developer (GH-600). Covers all 6 exam objectives with topic checklists, quick notes, and practice links.

Download PDF Notes

Exam Overview

Key facts before you start studying

Duration

120 min

Questions

Exam Code

GH-600

Domains

Quick Navigation

Jump to a specific exam domain

1. Prepare Agent Architecture and SDLC Processes (17%)
2. Implement Tool Use and Environment Interaction (23%)
3. Manage Memory, State, and Execution (13%)
4. Perform Evaluation, Error Analysis, and Tuning (17%)
5. Orchestrate Multi-Agent Coordination (17%)
6. Implement Guardrails and Accountability (13%)

Domain 1

17%

Prepare Agent Architecture and SDLC Processes

Integrate agents into SDLC, define planning/execution boundaries, configure observability

Topics to Master

Defining agent inputs, outputs, and success criteria
Integrating agents into pull request and CI workflows
Separating agent planning phase from execution phase
Configuring agent observability and durable audit artifacts
Least-privilege access and graduated autonomy models
Pre-execution plan validation and approval gates

Domain 2

23%

Implement Tool Use and Environment Interaction

Configure tools and permissions, MCP servers, CI integration, error handling, rollbacks

Topics to Master

Selecting and configuring agent tools with minimum required permissions
Configuring MCP servers and the MCP allow list
Configuring GitHub remote MCP servers
Scoping agents to specific repositories and branches
Invoking agents within CI workflow steps
Error handling: retry, rollback, and escalation paths
Traceability: recording tool calls with parameters and results

Domain 3

13%

Manage Memory, State, and Execution

Memory types, context drift prevention, state persistence, cross-agent state sharing

Topics to Master

Short-term vs. long-term vs. external memory
Scoping memory to task-relevant information
Memory expiration, pruning, and reset rules
Capturing task progress as durable artifacts
Detecting and correcting context drift
Preventing conflicting and stale context in multi-agent systems

Domain 4

17%

Perform Evaluation, Error Analysis, and Tuning

Define success criteria, analyze failures, classify root causes, tune agent behavior

Topics to Master

Specifying quantitative and qualitative evaluation signals
Automated scanning tools for evaluation signal generation
Root cause categories: reasoning errors, tool misuse, context/environment issues
Analyzing agent failures using logs, plans, traces, and workflow artifacts
Tuning instructions, memory usage, and tool access based on evaluation
Preventing evaluation set overfitting

Domain 5

17%

Orchestrate Multi-Agent Coordination

Orchestration patterns, conflict detection, multi-agent observability, failure recovery

Topics to Master

Pipeline, fan-out/fan-in, orchestrator-worker, and consensus patterns
Configuring agent isolation for parallel execution
Detecting and resolving overlapping changes and contradictory outputs
Configuring observability with correlation IDs and per-agent decision logs
Multi-agent recovery patterns: rollback and human-in-the-loop
Managing agent lifecycle: adding, updating, and retiring agents

Domain 6

13%

Implement Guardrails and Accountability

Autonomy levels, guardrail design, least privilege, human-in-the-loop, audit logging

Topics to Master

Classifying actions by risk: reversibility, impact scope, compliance sensitivity
Assigning autonomy levels to maximize delivery speed within compliance
Implementing guardrails and human-in-the-loop workflows
Least-privilege permissions and execution contexts
Blocking actions that violate security, compliance, or Responsible AI policies
Prompt injection defense through system-level permission restrictions

Continue Your Prep

Free CertStud resources for GH-600

Study Notes Exam Study Plan Start Practice Test

Back to Agentic AI Developer (GH-600)

Agentic AI Developer (GH-600) Study Guide

Domain-by-domain study guide for GitHub Certified: Agentic AI Developer (GH-600). Covers all 6 exam objectives with topic checklists, quick notes, and practice links.

Download PDF Notes

Exam Overview

Key facts before you start studying

Duration

120 min

Questions

Exam Code

GH-600

Domains

Quick Navigation

Jump to a specific exam domain

1. Prepare Agent Architecture and SDLC Processes (17%)
2. Implement Tool Use and Environment Interaction (23%)
3. Manage Memory, State, and Execution (13%)
4. Perform Evaluation, Error Analysis, and Tuning (17%)
5. Orchestrate Multi-Agent Coordination (17%)
6. Implement Guardrails and Accountability (13%)

Domain 1

17%

Prepare Agent Architecture and SDLC Processes

Integrate agents into SDLC, define planning/execution boundaries, configure observability

Topics to Master

Defining agent inputs, outputs, and success criteria
Integrating agents into pull request and CI workflows
Separating agent planning phase from execution phase
Configuring agent observability and durable audit artifacts
Least-privilege access and graduated autonomy models
Pre-execution plan validation and approval gates

Domain 2

23%

Implement Tool Use and Environment Interaction

Configure tools and permissions, MCP servers, CI integration, error handling, rollbacks

Topics to Master

Selecting and configuring agent tools with minimum required permissions
Configuring MCP servers and the MCP allow list
Configuring GitHub remote MCP servers
Scoping agents to specific repositories and branches
Invoking agents within CI workflow steps
Error handling: retry, rollback, and escalation paths
Traceability: recording tool calls with parameters and results

Domain 3

13%

Manage Memory, State, and Execution

Memory types, context drift prevention, state persistence, cross-agent state sharing

Topics to Master

Short-term vs. long-term vs. external memory
Scoping memory to task-relevant information
Memory expiration, pruning, and reset rules
Capturing task progress as durable artifacts
Detecting and correcting context drift
Preventing conflicting and stale context in multi-agent systems

Domain 4

17%

Perform Evaluation, Error Analysis, and Tuning

Define success criteria, analyze failures, classify root causes, tune agent behavior

Topics to Master

Specifying quantitative and qualitative evaluation signals
Automated scanning tools for evaluation signal generation
Root cause categories: reasoning errors, tool misuse, context/environment issues
Analyzing agent failures using logs, plans, traces, and workflow artifacts
Tuning instructions, memory usage, and tool access based on evaluation
Preventing evaluation set overfitting

Domain 5

17%

Orchestrate Multi-Agent Coordination

Orchestration patterns, conflict detection, multi-agent observability, failure recovery

Topics to Master

Pipeline, fan-out/fan-in, orchestrator-worker, and consensus patterns
Configuring agent isolation for parallel execution
Detecting and resolving overlapping changes and contradictory outputs
Configuring observability with correlation IDs and per-agent decision logs
Multi-agent recovery patterns: rollback and human-in-the-loop
Managing agent lifecycle: adding, updating, and retiring agents

Domain 6

13%

Implement Guardrails and Accountability

Autonomy levels, guardrail design, least privilege, human-in-the-loop, audit logging

Topics to Master

Classifying actions by risk: reversibility, impact scope, compliance sensitivity
Assigning autonomy levels to maximize delivery speed within compliance
Implementing guardrails and human-in-the-loop workflows
Least-privilege permissions and execution contexts
Blocking actions that violate security, compliance, or Responsible AI policies
Prompt injection defense through system-level permission restrictions

Continue Your Prep

Free CertStud resources for GH-600

Study Notes Exam Study Plan Start Practice Test