Guide · Agent skill
How to use the Exam Forecast agent skill (exam-forecast)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The exam-forecast skill is an agent skill for law students. When you give it several past exams from the same professor, it looks for repeating patterns and uses them, together with your current syllabus, to suggest where to spend review time. This guide explains what the skill does, how to install it, how to trigger it, what it needs and where its limits are, so you can use it without expecting it to predict the actual exam.
Exam Forecastanthropics/claude-for-legal · Apache-2.0 Download the skill Quick start
- 1Download the ZIP. Get the exam-forecast ZIP from the directory page and unzip it.
- 2Place the skill folder. Copy the unzipped folder into ~/.claude/skills/, or into a project's .claude/skills/ folder.
- 3Or upload it in the Claude apps. Go to Settings → Capabilities → Skills and upload the skill there.
- 4Gather past exams. Collect past exams from the same professor as PDFs, pasted text or file paths.
- 5Make a natural request. Ask for the class by name, for example "Analyze these past Contracts exams and forecast emphasis for the final."
- 6Review the forecast. Read the saved forecast, check the sample confidence and caveats, and use the weights to plan review time.
What is an agent skill, and what does exam-forecast make the agent do?
An agent skill is a folder containing a SKILL.md file with instructions that an AI agent follows when a matching request comes in. exam-forecast tells the agent how to analyze past exams from the same professor.
The agent reads each past exam for its format, subject coverage, question style, fact-pattern density and recurring traps, then compares the exams to find what stays stable and what varies. It then combines those patterns with your current syllabus to produce a forecast of subject weights, likely question format and study emphasis.
- Subject weighting across past exams
- Question formats and call styles used by the professor
- Traps the professor tends to hide in fact patterns
- How often policy questions appear compared with doctrine
- A study emphasis recommendation: heavy, moderate and sanity-check topics
How do you install the exam-forecast skill?
Download the ZIP for the skill from the directory page and unzip it. For a personal install, place the unzipped folder in ~/.claude/skills/. To scope it to one project, place it in that project's .claude/skills/ folder instead.
You can also upload the skill in the Claude apps under Settings → Capabilities → Skills.
- Personal install: ~/.claude/skills/
- Project install: the project's .claude/skills/ folder
- Claude apps: Settings → Capabilities → Skills, then upload
~/.claude/skills/How do you trigger the skill with a natural request?
The skill is triggered by requests such as "what's on the exam", "analyze past exams", "predict the exam", or by sharing past exams in the conversation. You can also name the class and attach or point to the past exams.
The skill's argument hint asks for a class name, with past exams shared or paths to them. A clear request that names the class and includes the exams works best.
- "Analyze these three past exams from Professor [name] for Contracts and tell me where to focus."
- "Here are my past exams for Torts. Find the patterns and forecast emphasis for the final."
- "Analyze past exams" followed by the PDFs or file paths
What do you need before you start?
You need past exams from the same professor, supplied as PDFs, pasted text or file paths. The skill works best when every exam is from the same professor, because same-professor exams are the strongest signal.
Two inputs are optional. A class context file at ~/.claude/plugins/config/claude-for-legal/law-student/CLAUDE.md can hold class, professor, exam format and syllabus details. A current-class syllabus lets the skill compare what has been covered with what has been tested.
- Required: past exams as PDFs, pasted text or file paths
- Optional: class context file at ~/.claude/plugins/config/claude-for-legal/law-student/CLAUDE.md
- Optional: current-class syllabus
What does the forecast look like?
The skill writes a forecast in a fixed structure. It opens with the header STUDY NOTES — NOT LEGAL ADVICE, then gives the number of past exams analyzed, a sample confidence rating (thin under 3, moderate 3 to 5, strong 6 or more), and any caveats such as an open-book exam being compared with a closed-book one.
The body includes a historical subject weighting table, a question-style forecast, professor hobby horses to watch, topics covered but rarely tested, and a study emphasis split between heavy, moderate and sanity-check topics. The skill saves the forecast as a dated Markdown file under ~/.claude/plugins/config/claude-for-legal/law-student/exam-forecasts/[class]/. It is versioned, so re-running after a new past exam adds a new dated file.
- Heavy topics: about 40-50% of study time
- Moderate topics: about 30-40%
- Sanity check: about 10-20%, for topics covered but historically under-represented
What are the safety notes: scripts and network access?
According to the directory listing, this skill runs no scripts and uses no network access. Its work is reading the exams you provide and writing its output file.
Your past exams are read from the files or text you supply, and the skill writes its forecast to the location described above. Check the output file and your own copies of the exams, since they are the source material.
- Runs scripts: no
- Uses network: no
- Reads: the past exams you supply and the optional class context file
- Writes: the dated forecast Markdown file in the exam-forecasts folder
What people use it for
- Deciding how many hours to give consequential damages versus formation before a Contracts final, based on past exam weightings
- Spotting a professor who regularly hides the jurisdictional issue in otherwise clean facts, and practising for it
- Comparing exam styles across a course in a study group
- Choosing between issue-spotting practice and short-answer practice based on the usual question format
- Checking whether policy questions appear often enough to warrant reading policy material closely
Frequently asked questions
Is exam-forecast free to use?
The source material for this listing does not state pricing or licence terms beyond the Apache-2.0 licence noted for the repository. Check the licence file in the repository for the terms that apply to the code.
How do I install exam-forecast?
Download the ZIP from the directory page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. You can also upload it in the Claude apps under Settings → Capabilities → Skills.
What do I need to use exam-forecast?
You need past exams from the same professor as PDFs, pasted text or file paths. A class context file and a current-class syllabus are optional but help the forecast.
How many past exams do I need?
The skill warns that patterns from one or two exams are noise. It rates a sample of under 3 exams as thin, 3 to 5 as moderate and 6 or more as strong.
Will exam-forecast tell me what will be on my exam?
No. The skill states that it cannot tell what will be on the exam. It shows what has appeared on past exams and what is likely to recur, and it frames the output as a weighting guide for study time.
What should I do if my professor is new and has no past exams?
The skill cannot forecast without past exams. It can only list the subjects covered in the syllabus, so use it for that and plan your study with other materials.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use exam-forecast if you have several past exams from the same professor and want a structured way to decide where to spend review time. Skip it if you have one or two exams, a new professor with no past exams, or if you expect it to predict the actual questions, since the skill itself says it cannot do that.
Download the skill