AgentPrep
Get AgentPrep
§ 01Opening

Know the day you'll be ready

You don't need more content.
You need a stopping point.

Every course, PDF and playlist is built to keep you going. AgentPrep is built to tell you when to stop — a daily quest on Telegram, real tutoring in Claude Code, and one readiness score, per domain, that runs the arithmetic and hands you a date.

CCAR-F, CCAO-F, CCDV-F and CCAR-P prep is available now, same subscription, no extra charge. Works in English, Português and Español.

§ 02Method

The method

Three pieces. One score.

Study where you already live. The bot and the skill are two doors into the same progress — neither one keeps score on its own.

Telegram bot

A quest every morning: five scenario questions, answered where you already are. Streaks and XP keep it a habit instead of a chore. Talk to it — "what's my level," "focus on MCP" — no command syntax to memorize.

Claude Code skill

When five questions aren't enough, go deep: Socratic tutoring that explains the why, plus full 60-question, 120-minute mock exams. It runs on your own Claude subscription, so the tutoring costs us nothing — and costs you nothing extra.

npx skills add marcelomar21/agentprep-skill

Readiness score

The number the other two feed. Per-domain mastery, a star rating, and a forecasted ready date — recomputed after every single answer. The full arithmetic is below; nothing here is hidden from you.

§ 03Readiness

No black box

The arithmetic of ready.

This is the whole model — what the bot and the skill actually compute after every answer. No hidden weighting, no vibes.

Mastery = round( 100 × accuracy1 × coverage2 )
★ Stars = round( mastery ÷ 20 )0 to 5
Ready = min( mastery, every domain ) ≥ 803
Ready date = today + ⌈ Σ questions to ready ÷ pace4
  1. Accuracy is an exponentially-weighted average of your answers in that domain — half-life of 20 questions, so what you got wrong three weeks ago counts for less than what you got wrong this morning.
  2. Coverage is distinct questions you've actually seen, divided by the domain's bank, capped at 1. Getting the same five questions right on repeat does not move this number.
  3. Every domain clears 80 on its own. The overall average can look fine while one domain quietly isn't — this refuses to let that hide.
  4. Pace is your stated daily goal, or your average over the last 7 active days if you haven't set one. The projection assumes your accuracy keeps climbing toward 90% as you go — it's a forecast, not a promise, and it moves every time you answer.

No LLM grades you behind the scenes. Every number above, you could compute yourself from your own answer log.

Animated demo of the AgentPrep Telegram bot: asking "what's my level?" returns a readiness card with stars per domain, asking "when will I be ready?" returns a forecasted date, then a real practice question is answered for XP.
Fig. 1 — The forecast, live in the chat. Ask "when will I be ready?" and get a date, not a guess.
Mastery forecast crossing the readiness threshold Illustrative line chart. A mastery score climbs from about 38 to 66 over three weeks of practice, plotted as a solid line. A dashed projection then continues the same trend forward and crosses the 80-point readiness line on August 9 — the computed ready date. 0 20 40 60 100 Ready — 80 3 wks ago Today Aug 9 +4 wks
Fig. 2 — Example forecast. Real accounts start from zero.
§ 04Certifications

Certifications

Pick a start. Switch for free.

One subscription. All the prep material lands in it — for today's exams and for every certification Anthropic launches next.

  • 01 CCAR-F— Claude Certified Architect, Foundations (formerly CCA-F) ● Available now
  • 02 CCAO-F— Claude Certified Associate, Foundations (opened July 13) ● Available now
  • 03 CCDV-F— Claude Certified Developer, Foundations ● Available now
  • 04 CCAR-P— Claude Certified Architect, Professional ● Available now

Every new Anthropic certification: included, same subscription.

§ 05Pricing

Pricing

One number. No upsells.

You already read the method and the math. This is what it costs.

The Offer

$49.90per year

  • All four certifications — every new one after, at no extra cost
  • Telegram bot + Claude Code skill, one shared score
  • English, Português and Español
  • A living question bank, updated continuously
  • Your readiness score and forecasted ready date
Get AgentPrep

Secure checkout, Merchant-of-Record billing. License key by email, refunds on request.

§ 06FAQ

FAQ

Before you ask.

Q. Do I need a Claude subscription?
A.

The Claude Code skill runs on your own Claude subscription — Pro, Max, or an API key — so the deep tutoring costs us nothing, and costs you nothing extra. The Telegram bot needs no Claude subscription at all; it runs entirely on its own.

Q. Are the questions from the real exam?
A.

No — and they can't be. Every question is 100% original and scenario-based, written from public material: official exam guides, Anthropic docs, Academy content. Using real exam questions would violate Pearson VUE's NDA. We don't, on principle.

Q. What's a "living question bank"?
A.

It's continuously updated to track the exam and the Claude ecosystem as both change — not a fixed pile you empty out. Readiness isn't "you cleared the bank," it's your mastery holding 80+ in every domain. That's what §03 actually measures.

Q. Can I switch certifications?
A.

Anytime, at no cost. One subscription covers every certification we support — pick one to start, switch whenever you want, and new ones join the same subscription the day they ship.

Q. How do I get my license, and can I get a refund?
A.

A license key lands in your inbox right after checkout — paste it into the bot or the skill, either one activates both. Refunds: ask, and we sort it out.

§ 07Close

Know the date. Then go take the exam.

$49.90 a year. Every Anthropic certification. We're always expanding — new ones are included in your subscription as they launch.

Checkout opens soon

Message @AgentPrep_bot on Telegram — we'll notify you the moment it's live.