Agentic AI Certification Hub is a structured learning and reference repository for agentic AI concepts, Claude, GitHub, architecture foundations, practice questions, and cross-certification study.
This repo exists to separate durable agentic AI architecture concepts from vendor-specific exam vocabulary. The core material covers foundations such as orchestration, context, tools, memory, guardrails, and evaluation. The certification folders then apply those foundations to specific platforms and exam surfaces.
- Architecture practitioners preparing for agentic AI certifications.
- Enterprise and solution architects who need a practical grounding in agentic AI systems.
- Builders using Claude, GitHub Copilot, MCP, and AI-assisted delivery workflows.
- StudioSix and Velocity Architecture practitioners who need a learning layer for agentic architecture concepts.
This repository supports the Velocity Architecture / StudioSix ecosystem by strengthening the AI-assisted delivery learning layer. It is not the core framework and it is not a product repo; it is a study and reference hub that helps practitioners understand the concepts behind agentic architecture work.
Usable study repository. Current content covers Claude Certified Architect foundations and GitHub agentic AI certification preparation, with shared cross-certification foundations in core/.
| Certification | Issuer | Exam | Status | Domain focus |
|---|---|---|---|---|
| Claude Certified Architect – Foundations | Anthropic | CCA-F | Live · March 2026 | Claude API, Claude Code, MCP |
| GitHub Certified: Agentic AI Developer | GitHub / Microsoft | GH-600 | Beta · May 2026 | GitHub Copilot, SDLC, Copilot agents |
agentic-ai-certification/
│
├── README.md ← This file
│
├── core/ ← Platform-agnostic foundations
│ ├── Agentic-Concepts.md ← Orchestration, context, tools, memory
│ ├── Anti-Patterns-Master.md ← All failure modes, vendor-neutral
│ └── Mental-Models.md ← Diagnostic reasoning framework
│
└── certifications/
│
├── CCA-F/ ← Anthropic · Claude Certified Architect
│ ├── README.md ← CCA-F overview and domain map
│ ├── domain-breakdowns/ ← D1–D5 deep dives
│ ├── study-guides/ ← Quick ref, anti-patterns, field guide
│ ├── practice-questions/ ← 57 scenario questions across all domains
│ ├── prompt-engineering/ ← Tutor prompts, system prompt templates
│ └── resources/ ← Links, schedule, fellowship research
│
└── GH-600/ ← GitHub · Agentic AI Developer
├── README.md ← GH-600 overview and domain map
├── domain-breakdowns/ ← D1–D6 deep dives
├── study-guides/ ← Quick ref, anti-patterns
├── practice-questions/ ← Scenario questions per domain
└── resources/ ← Links, beta discount, study schedule
New to agentic AI architecture: Start in core/. Read Agentic-Concepts.md, then Mental-Models.md. These are the foundations every certification exam tests from different angles.
Targeting CCA-F first: Go directly to certifications/CCA-F/README.md. Everything is built and ready.
Targeting GH-600 first: Go to certifications/GH-600/README.md. Study core/ alongside it — GH-600 is GitHub-surface agentic architecture, and the core concepts are identical.
Studying both simultaneously: Read core/ once. Then use each certification folder for platform-specific vocabulary, tooling, and exam scenarios. The overlapping domains reinforce each other.
The same architectural concepts appear in both exams under different names and through different tooling.
| Concept | CCA-F Domain | GH-600 Domain |
|---|---|---|
| Agent loop and stop reason handling | D1 · Agent Architecture (27%) | D1 · Agent Architecture & SDLC (15–20%) |
| Multi-agent orchestration | D1 · Agent Architecture (27%) | D5 · Multi-Agent Coordination (15–20%) |
| Tool design and routing | D4 · Tool Design & MCP (18%) | D2 · Tool Use & Environment (20–25%) |
| Context and memory management | D5 · Context & Reliability (15%) | D3 · Memory, State & Execution (10–15%) |
| Guardrails and enforcement | D1 · Hooks vs prompts | D6 · Guardrails & Accountability (10–15%) |
| Evaluation and error analysis | D3 · Prompt Engineering (20%) | D4 · Evaluation, Error & Tuning (15–20%) |
| CI/CD and SDLC integration | D2 · Claude Code (20%) | D1 · Agent Architecture & SDLC (15–20%) |
Study pattern: Learn the concept in whichever exam you're tackling first. When you move to the second exam, you're learning the platform surface — not the concept again.
The first 100 people who take Exam GH-600 (beta) on or before May 31, 2026 can get 80% off market price. Use discount code GH600Flanders at registration. Spots are limited and first-come, first-served. The beta exam is not available in Turkey, Pakistan, India, or China.
Australia is eligible. Register via Pearson VUE through Microsoft Learn.
Phil Myint · ZenCloud Global Consultants · May 2026