Claude, Exam Forecast — What an Exam-Forecast Skill Actually Promises

Does a Claude skill called exam-forecast mean it has somehow seen your actual exam? It hasn't, and it can't. A professor's exam file is locked until exam day. What the skill reads instead is everything already public about the course: the syllabus, the reading list, and past exams. From that, it builds a ranked list of topics to guide where you start studying.

The problem it's solving

Final exam season compresses weeks of reading into a few days. Rereading everything from the first page forward means running out of time before reaching the end. If a contracts final is three weeks out, there needs to be a way to prioritize which topics deserve attention first, without simply guessing.

What the skill actually reads

Feed it a syllabus and past exams for the course, and exam-forecast hands back a ranked list of topics. The natural assumption is that this means Claude has somehow seen this year's actual exam and can hand over the real questions early. It has not, and it cannot. A professor's exam file stays locked until exam day. Instead, the skill works from the pattern already sitting in public course material: what topics the syllabus emphasizes, what the reading list covers, and how often each topic has come up on past exams for that course.

How the ranking gets built

A skill is a folder Claude reads before it works, and its SKILL.md file lists the steps in a fixed order: gather the sources, count how often each topic was tested before, then rank what's left. The output is a probability read from a public pattern, not a leak of confidential material.

Where it breaks down

This only works where a pattern already exists to count. A brand-new course, or a professor who rewrites the exam format every year, leaves nothing to count. No history means no forecast; the skill has nothing to rank against.

Reading the ranked list correctly

Applied to a contracts final, the ranked list might come back with consideration doctrine first, having been tested in four of the last five years, followed by the statute of frauds, then narrower exceptions. But ranked first does not mean guaranteed. Consideration doctrine could still sit out this particular year, the way any pattern can break once. And ranked low does not mean safe to skip. A topic near the bottom of the list can still be the one curveball that actually appears on the page. The skill re-orders study time; it never deletes a topic from consideration.

What it promises and what it doesn't

Exam-forecast does not show you the exam. It shows you where the class has historically leaned, so you know what to study first when time is limited. That is a study-order tool built from public patterns, not a prediction guaranteed to match what actually appears.

Key takeaways

• Exam-forecast never sees the actual exam; it reads the syllabus, reading list, and past exams, all public course material.
• It counts how often each topic has appeared on past exams and ranks accordingly.
• The skill only works when there is a pattern to count; a new course or a professor who changes format yearly leaves no forecast.
• Ranked first is not a guarantee, and ranked last is not safe to skip.
• The output re-orders where you start studying; it never removes a topic from your list.

Who this is for

Students preparing for exams with a syllabus and access to past exams for the course, who want a way to prioritize study time under real time pressure without mistaking a ranked list for a leaked exam.