How I Scored 910/1000 on CCAF Claude Certification and Open-Sourced My Last Cheat Sheet

Ahmed Hasan — Tue, 28 Jul 2026 10:25:01 +0000

I passed Claude Certified Architect Foundations (CCAF) with 910/1000 on 15 July 2026, and I'm the first at my company to finish it. I open-sourced the study notes I built: github.com/ahmdhsn-dev/ccaf-study-guide.

CCAF is scenario-heavy: you pick the right architecture when an agent forgets context, a generic tool keeps misfiring, or a customer escalation lands mid-process. The official guide covers all of it, but it's dense. The repo distills 30 exam tasks across 5 domains into two complementary views: what to do (patterns and traps) and what not to do (anti-patterns).

What's inside: a one-page concept map, a trap taxonomy (why wrong answers sound right), and an anti-patterns catalog that maps every "avoid / do not / instead of" warning from the official guide. If a distractor sounds plausible, check the traps first; if it sounds like a shortcut the guide rejects, check the anti-patterns.

Use it as your last cheat sheet, not your textbook. Take a mock first to see where you stand, then study the official guide and use the repo to close the gaps.

Start here

Get access via Anthropic Academy / Skilljar
Take a free mock (CyberSkill or Certification Guide) to baseline where you stand
Read the official exam guide at least 3 times (I printed mine to cut digital distractions)
Practice on a real project alongside study: build an agent loop, wire up MCP tools, or configure Claude Code. Scenarios stick better when you've lived them.
Retake mocks until you hit 900+, then book the real exam

Full workflow: study/README.md

Recommended Academy path

Building with the Claude API → Claude Code in Action → Introduction to Agent Skills → Introduction to MCP

I completed all four with hands-on labs in May 2026. Production work with skills, MCP servers, and tool orchestration helped, but the exam tests guide patterns, not how your repo is wired.

Study path (4 docs)

Step	Doc	What you get
1	Concept map	All 30 tasks on one page
2	Mock exam trap guide	Why wrong answers sound right
3	Anti-patterns catalog	What the guide says not to do
4	API & product FAQ	Exam vs production gotchas

Domain weights: Agentic Architecture (27%) · Tool Design & MCP (18%) · Claude Code Config (20%) · Prompt & Structured Output (20%) · Context & Reliability (15%). See the concept map for task breakdown.

Mock exam strategy

Take the CyberSkill mock first (closest to the real exam), then the Certification Guide short mock. My first CyberSkill run scored 850/1000; after mapping misses to the guide and retaking, I was consistently 950+ before booking. For every miss, map it to the trap taxonomy, cross-check the anti-patterns catalog, re-read the matching section in the official guide, and repeat until you're consistently at 900+ (I booked at 950+).

Mocks are your prerequisite gate, not an arbitrary iteration cap. You won't feel ready. The exam has stop_reason: end_turn before you do, and that's normal.

Nine misses on my first CyberSkill run; these three patterns accounted for most of them (full mapping in the cheat sheet below):

Context (Task 5.1): Progressive summarization beats a "structured issue data layer" when context fills up
Tool design (Task 2.1): Split purpose-specific tools beat a generic analyze_document with an enum mode
Escalation (Tasks 1.4, 5.2): Structured handoff brief, not a reference ID; angry + no context → acknowledge + one question

Trap cheat sheet

High-yield patterns from mock exams. Full taxonomy: 02-mock-exam-trap-guide.md.

Context filling up (multi-issue)     → 5.1 Summarize resolved; active thread full
Generic tool, inconsistent output    → 2.1 Split typed tools
Messy source formats                 → 4.3 Schema + normalization in prompt
Function behind wrappers             → 2.5 List aliases → grep each
Agent forgets mid-conversation       → 5.1 Pass full messages array
Unknown bug path                     → 1.6 Adaptive investigation
Long code exploration (30+ min)      → 5.4 Scratchpad file
Mid-process escalation               → 1.4 Structured handoff brief
Angry + no context                   → 5.2 Acknowledge + one question
Identity multi-match                 → 5.2 Ask identifiers, never guess

Trap	Sounds like...	Guide-preferred
Sounds enterprise	Structured data layer, reference IDs	Progressive summarization
Sounds efficient	Re-fetch via MCP, parallel workers	Trim/summarize context
Sounds smart	Confidence threshold, routing classifier	Explicit criteria; programmatic gates
Sounds helpful	Immediate escalation, unauthorized retry	Calibrated: brief, one question, or resolve
Sounds thorough	Plan everything first	Adaptive evidence-driven steps
Sounds simple	Better description on generic tool	Split tools + typed schemas
Sounds pragmatic	Post-processing normalization	Canonical format at extraction

Top 5 anti-patterns (high exam yield)

From 03-anti-patterns-catalog.md:

1. Heuristic customer selection on multi-match: ask for email, phone, or order ID; never confidence-threshold auto-pick (Task 5.2).

2. Sliding window drops resolved threads: progressive summarization on resolved issues; keep active thread full (Task 5.1).

3. Generic analyze_document + free-text mode: split into extract_data_points, summarize_content, etc. (Task 2.1).

4. Batch API for pre-merge blocking checks: sync Messages API for blocking; Batch for overnight audits (Task 4.5).

5. Giving synthesis agent web search tools: scope to verify_fact; route new lookups through coordinator (Task 2.3).

Exam vs production (quick hits)

Topic	Exam (guide)	Production (2026)
Structured output	`tool_use` + JSON schema	Also `output_config.format`; prefill removed on new models
Batch tool use	No multi-turn in one request	Server tools + `pause_turn`; client MCP loops still multi-request
`/memory`	Diagnostic command (in scope)	Plus expanded CLAUDE.md / rules / skills ecosystem

Full breakdown: 04-api-and-product-faq.md

Repo link

github.com/ahmdhsn-dev/ccaf-study-guide. MIT licensed, unofficial, not affiliated with Anthropic.

Community links and extra mocks: resources/links.md

Disclaimer: Based on the exam taken 15 July 2026. Unofficial, not affiliated with Anthropic. Verify latest exam requirements before you sit.

Is working remote from Skardu possible?

Ahmed Hasan — Mon, 13 Jun 2022 15:09:28 +0000

Introduction

While doing my undergraduate study in Computer Science, I made a lot of friends that are now doing remote jobs or are working remotely from home. A bunch of us decided to work from a beautiful natural place where our eyes can relax away from digital screens without affecting our work.

About the Place

Skardu is located in the Gilgit-Baltistan (northern region) of Pakistan. The city is an important gateway to the important nearby mountains of the Karakoram mountain range. It is a known hub for mountaineers to stay before expediting to reach one of the highest peaks in the world.

Is work from Skardu possible?

Following are my findings on the significant three factors for smooth remote work.

Electricity
Mobile Network
Internet

Electricity
Since the main source of electricity production is from hydroelectric plants installed in rivers and canals. Visiting the place in summer will give you a constant availability of electricity. You may need more power backup if visiting out of season. Scheduled load shedding will be common close to winters. The ideal time to get good electricity would be from June till August.

Mobile network
The mobile networks of Pakistan like Ufone and Zong works well for calls and messages but will not work for internet data connectivity. There is a different mobile network named SCOM in Skardu. For internet, you can get a SCOM sim at a very decent price and can have a high-speed internet package on it as well.

Internet
You can get high-speed internet in the main city areas.
We were able to get a decent enough speed at Skardu main bazaar.
The speed in Shigar actually exceeded the speed I get in my home. Overall, I would say that the internet will not be an issue as long as electricity is not an issue. Having a good internet package can really make it possible to work without any hassle.
Speed in Skardu city

Speed in Shigar

Did we actually work?

While planning and evaluating the available options, our final decision was to take the time off to clear our heads (yes we need that a lot) instead of working remotely.
We took annual leaves for a week as we were not too sure about the limitations that we might face for smooth remote work, and postponed that remote work plan for later.
Some of us did however take our workstations to move images and stuff from DSLR (thank you Abbas for letting me borrow it) and Slack communications (yeah! not a good idea for a vacation).
Other than image editing, migration, and communications on Slack, I did not work on anything related to development. Nouman however was able to find time and work on some side tasks without any hassle.

Our vacation summary video

If you are interested in knowing a bit more about the place and us, you can check out this Youtube video

Conclusion

Based on the three main factors we observed during our 7 days visit, we think it is very possible to work remotely from the beautiful valley of Skardu.

DEV Community: Ahmed Hasan