GH-600 Week-by-Week Learning Path¶
The Scrollmaster's Prescribed Curriculum — for those who prefer a map over wandering.
Total estimated time: ~40 hours over 6 weeks (~7 hours/week) Assumed background: Comfortable with GitHub Actions, basic Python or JS, VS Code, and basic AI-tool usage.
Choose your pace: 6-week standard track (default) · 2-week sprint (experience required) · micro-learning track (busy schedules)
Before You Begin¶
- Read the Skills Measured breakdown and mark which sub-skills you already know.
- Skim the Recommended Resources list.
- Set up your sandbox: clone this repo and run
./work/gh-600/setup.shto get the starter kit. - Ensure you have a GitHub account with Copilot access (free trial is fine for most quests).
Week 1 — Domain 1: Agent Architecture & SDLC (≈7 hrs)¶
Goal: Understand what an agent is in the SDLC, how planning differs from action, and how to make agent behavior observable.
| Day | Activity | Time |
|---|---|---|
| Mon | Read Skills Measured — Domain 1 · Skim Copilot coding agent docs | 1 hr |
| Tue | Complete Q1: Initiation Rites — SDLC Integration | 90 min |
| Wed | Complete Q2: The Three Sigils — Plan vs. Action Boundaries | 90 min |
| Thu | Complete Q3: The All-Seeing Eye — Observability & Control | 90 min |
| Fri | Write a 1-page reflection: anti-patterns you've seen in your own AI usage; review Chapter I — Agents in the SDLC | 30 min |
Week 2 — Domain 2: Tool Use & Environment Interaction (≈8 hrs)¶
Goal: Configure and operate agents with real tools — including spinning up an MCP server and integrating into a CI workflow.
| Day | Activity | Time |
|---|---|---|
| Mon | Read MCP Quick Reference · Skim MCP specification | 1 hr |
| Tue | Complete Q4: Forging the Arsenal — Tool Selection & Permissions | 90 min |
| Wed | Complete Q5: The MCP Conclave — MCP Server Mastery | 2 hrs |
| Thu | Complete Q6: Bind the Realm — Dev Environment Integration | 2 hrs |
| Fri | Complete Q7: The Shield of Retries — Safe Execution & Error Handling | 2 hrs |
Week 3 — Domain 3: Memory, State & Execution (≈5 hrs)¶
Goal: Design durable agent memory strategies and detect/correct context drift.
| Day | Activity | Time |
|---|---|---|
| Mon | Read Skills Measured — Domain 3 | 30 min |
| Tue | Complete Q8: Vaults of Recollection — Memory Strategies | 90 min |
| Wed | Complete Q9: Anchoring the Drifting Agent — State Persistence & Drift | 90 min |
| Thu | Complete Q10: Crossing the Tool Planes — State Continuity Cross-Tools | 90 min |
| Fri | Review Chapter III — Vaults of Recollection: Memory & State | 30 min |
Week 4 — Domain 4: Evaluation, Error Analysis & Tuning (≈6 hrs)¶
Goal: Define testable success criteria, read agent failure traces, and tune instructions/tools based on evidence.
| Day | Activity | Time |
|---|---|---|
| Mon | Read Evaluation Signals Table | 30 min |
| Tue | Complete Q11: The Oracle's Rubric — Success Criteria & Signals | 90 min |
| Wed | Complete Q12: The Necromancer's Inquest — Failure Root Cause Analysis | 90 min |
| Thu | Complete Q13: Reforging the Agent's Mind — Behavior Tuning | 2 hrs |
| Fri | Review Chapter IV — The Oracle Rubric: Evaluation & Tuning | 30 min |
Week 5 — Domain 5: Multi-Agent Coordination (≈7 hrs)¶
Goal: Orchestrate parallel agents, capture audit trails, recover from failures, and manage agent lifecycles.
| Day | Activity | Time |
|---|---|---|
| Mon | Read Skills Measured — Domain 5 | 30 min |
| Tue | Complete Q14: The Council of Many — Orchestration Patterns | 2 hrs |
| Wed | Complete Q15: The Scribe's Codex — Multi-Agent Observability | 90 min |
| Thu | Complete Q16: When Familiars Fall — Failure & Recovery | 90 min |
| Fri | Complete Q17: The Agent Pantheon — Lifecycle Management | 2 hrs |
Week 6 — Domain 6 & Capstone (≈7 hrs)¶
Goal: Apply Responsible AI policy, classify risk, implement HITL gates, then prove mastery in the capstone.
| Day | Activity | Time |
|---|---|---|
| Mon | Read Autonomy Levels Matrix | 30 min |
| Tue | Complete Q18: The Autonomy Scales — Autonomy Levels Matrix | 90 min |
| Wed | Complete Q19: The Warden's Pact — Guardrails & HITL | 2 hrs |
| Thu | Review Skills Checklist — mark every box | 30 min |
| Fri–Sat | Complete 🏆 Capstone: Trial of the Agentic Codex (includes 50-question self-assessment) | 4 hrs |
Exam Day Checklist¶
- All 19 quests completed and validated
- Skills Checklist: every box ticked
- Capstone self-assessment score ≥ 80%
- Reviewed Glossary the night before
- Rested, fed, and ready
- Schedule your exam at Pearson VUE
2-Week Sprint¶
For experienced GitHub Copilot users who already understand agent concepts and need structured exam preparation rather than a full learning journey.
Prerequisites: Active GitHub Copilot usage, familiarity with GitHub Actions, and basic awareness of LLM concepts. Complete the Skills Checklist self-assessment first — aim to already pass ≥60% before starting the sprint.
Total time: ~20 hours over 14 days
| Day | Focus | Activities | Time |
|---|---|---|---|
| 1 | Domains 1–2 overview | Skim Skills Measured + Q1 + Q2 | 3 hrs |
| 2 | Domain 2 tools | Q4 + Q5 (MCP) | 3 hrs |
| 3 | Domain 2 deep dive | Q6 + Q7 | 3 hrs |
| 4 | Domain 3 memory | Q8 + Q9 + Q10 | 3 hrs |
| 5 | Domain 4 eval | Q11 + Q12 | 2 hrs |
| 6 | Domain 4 tuning | Q13 + review Chapter IV | 2 hrs |
| 7 | Rest day — light review of Glossary | 30 min | |
| 8 | Domain 5 part 1 | Q14 + Q15 | 3 hrs |
| 9 | Domain 5 part 2 | Q16 + Q17 | 3 hrs |
| 10 | Domain 6 | Q18 + Q19 | 3 hrs |
| 11 | Weaknesses | Re-do quests where you scored below 80% | 2 hrs |
| 12 | Reference pass | Read all 7 Quick-Reference Notes | 1 hr |
| 13 | Capstone attempt | Full capstone — identify gaps | 4 hrs |
| 14 | Final review | Gaps from capstone + Skills Checklist | 2 hrs |
Micro-Learning (15 min/day)¶
For those with demanding schedules who can only commit to short, daily sessions. Completion time: ~3 months at 5 days/week.
How it works: Each 15-minute session is a single, self-contained activity. Sessions build sequentially — do not skip. Mark each checkbox as you go.
Phase 1 — Orientation (Days 1–10)¶
| Day | Session |
|---|---|
| 1 | Read Exam at a Glance (the hub table) |
| 2 | Read Domain 1 Skills Measured |
| 3 | Skim Copilot coding agent docs |
| 4 | Start Q1 (first half — read through "Quest Setup") |
| 5 | Finish Q1 (complete all exercises) |
| 6 | Start Q2 (first half) |
| 7 | Finish Q2 |
| 8 | Start Q3 |
| 9 | Finish Q3 |
| 10 | Read Chapter I — Agents in the SDLC |
Phase 2 — Tools & Environment (Days 11–30)¶
Work through Q4–Q7 across these 20 sessions, reading reference material between quests. See the Week 2 schedule for the activity sequence — split each day's entry into 2–3 sessions of 15 minutes each.
Phase 3 — Memory, Evaluation & Coordination (Days 31–70)¶
Work through Q8–Q17 using the Weeks 3–5 schedules as a guide, splitting each day into micro-sessions.
Phase 4 — Guardrails & Capstone (Days 71–90)¶
Complete Q18–Q19, review all notes, and finish the capstone across the final 20 sessions.