GH-600: GitHub Agentic AI Developer
Browse certifications
Exam Resources
Official learning paths, exam details, skills measured, and community resources to supplement your study.
About the GH-600 Exam
Master agentic AI on GitHub — agent architecture and SDLC integration, tool use and MCP servers, memory and state, evaluation and tuning, multi-agent orchestration, and guardrails.
The complete practice exam for GitHub Certified: Agentic AI Developer (GH-600, ‘Developing in Agentic AI Systems’). Covers preparing agent architecture and SDLC processes, separating planning from execution and validating agent plans, selecting and permissioning agent tools, configuring MCP servers including the GitHub remote MCP server, registries and allow lists, scoping agents to repositories and branches and invoking them from CI, short-term, long-term and external memory with Copilot Memory, context drift and durable artifacts, success criteria, automated scanning and root-cause analysis, multi-agent orchestration, isolation, conflict resolution and recovery, and autonomy levels, least-privilege scoping and human-in-the-loop guardrails — every question a real-world scenario with full explanations.
Who Should Take This Exam?
The GH-600 is designed for professionals specialising in a focused domain. Prior certification or equivalent experience recommended.
Prerequisites: None (SDLC, GitHub workflow and Copilot experience recommended)
Typical study time: 6-10 weeks of focused study
Exam Quick Facts
| Detail | Value |
|---|---|
| Exam Code | GH-600 |
| Title | GitHub Agentic AI Developer |
| Duration | 120 minutes |
| Questions | ~60 |
| Pass Score | 700 / 1000 |
| Cost | $99 USD |
| Provider | Pearson VUE (GitHub Certifications) |
| Validity | 2 years |
| Prerequisites | None (SDLC, GitHub workflow and Copilot experience recommended) |
| Question Types | Multiple choice |
| Official Page | View on GitHub → |
Exam Domains & Weights
The GH-600 exam covers 6 domains. Focus your study time based on the weights below — higher-weighted domains have more exam questions.
| Domain | Weight | Practice Qs |
|---|---|---|
| Prepare agent architecture and SDLC processes | 17% | 44 |
| Implement Tool Use and Environment Interaction | 22% | 56 |
| Manage Memory, State, and Execution | 12% | 31 |
| Perform Evaluation, Error Analysis, and Tuning | 17% | 44 |
| Orchestrate Multi-Agent Coordination | 17% | 44 |
| Implement Guardrails and Accountability | 12% | 31 |
| Total | 100% | 250 |
💡 Study tip: Implement Tool Use and Environment Interaction carries the most weight (22%) — start there. Manage Memory, State, and Execution has the least (12%), but don’t skip it — exam questions can come from any domain.
Practice Exam — 250 Questions
Prepare for the GH-600 with our 250-question practice exam covering all 6 exam domains. Every question includes detailed explanations and maps to official exam objectives.
What you get:
- ✅ Exam simulation mode with timer
- ✅ Spaced repetition for weak areas
- ✅ Detailed explanations for every question
- ✅ Progress tracking across domains
- ✅ 20 free questions — no account needed
GitHub Certification Path
Start with GitHub Foundations (fundamentals), then specialise: Actions, Administration, Advanced Security, and the GitHub Copilot certification.
Related GitHub Certifications
If you’re studying for the GH-600, you might also be interested in these GitHub certifications:
- GH-900: GitHub Foundations — 250 practice questions
- GH-300: GitHub Copilot — 250 practice questions
- GH-200: GitHub Actions — 250 practice questions
- GH-100: GitHub Administration — 250 practice questions
- GH-500: GitHub Advanced Security — 250 practice questions
Study Tips
- Start with the heaviest domain — focus your time where the exam focuses its questions
- Use our practice exam — try the 20 free questions first to gauge your readiness
- Review explanations — don’t just check if you got it right; read why each answer is correct
- Simulate exam conditions — use the timed exam mode to practice under pressure
- Check the official page — official exam details always have the latest objectives