course-study-session
Skill2 starsChanged 56 days agoMITView on GitHub
Use when running a course study session: prep + gap-check.
In this repository16 other files it gives its agents.
- canvas-course-syncSkill
- categorizerSkill
- course-knowledge-baseSkill
- email-draftingSkill
- exam-prepSkill
- grade-calculatorSkill
- homework-checklistSkill
- meetings-and-clubsSkill
- note-formatting-rulesSkill
- photos-to-pdfSkill
- roadmapSkill
- school-advisorSkill
- study-notes-workflowSkill
- suite-updateSkill
- verification-before-completionSkill
- writing-for-agentsSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 15
Only then start homework: HW spans multiple lectures, and the review primes the problems.
- Distinctive: line 20
Notes often STOP mid-deck: after content-checking, list the deck's remaining sections the notes never reached — the uncovered stretch is …
- Very specific: line 31
Check answers carry units (engineering notation applies — 10 µA not 10).
- Distinctive: line 32
When the user flags a stale prerequisite ("years since I took calc"), embed a two-minute refresher at the point of use in a HW problem (e.g.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Course Study Sessions
Recurring job: user dedicates a session to a course (pre/post-lecture review, homework) and wants (a) confirmation he correctly understood the key ideas, (b) what he got wrong or missed vs. the professor's slides, (c) homework progress. Materials live under Desktop/Classes/<Course>/: slides in Materials/Lectures/ (numbered, e.g. <COURSE_CODE>-N_Topic), raw/processed notes in Notes/, HW sheets in Materials/Homework/HW{n}.pdf, textbook PDF in Materials/. Roadmap (Classes/Roadmap.md) has deadlines + load tags; course README has grading/AI policy.
Session flow (in order)
- Pre-read BEFORE the session starts (user is at lecture or away): extract the day's deck, pull the exact HW problems from the textbook, load Roadmap + README. Come in ready, not reading for the first time during review.
- Review previous lecture's notes vs. its deck (gap check below).
- Review today's notes vs. today's deck — same check.
- Only then start homework: HW spans multiple lectures, and the review primes the problems.
Gap check (slides = source of truth, notes = capture)
- Three outputs: what notes got right (confirm), wrong (correct in place), missed (list it).
- Notes often STOP mid-deck: after content-checking, list the deck's remaining sections the notes never reached — the uncovered stretch is the review target, not a footnote.
- Preserve every fact; never delete the user's
***clarify markers; don't solve his self-test exercises unless asked; equations in LaTeX. Full processing rules live in thenote-formatting-rulesskill — point future sessions there instead of duplicating.
Deck + textbook PDF handling
Lecture PDFs are image-heavy: read_file often fails with NeedsOcrError while a text layer exists on most pages. Extract with pymupdf to a temp .txt (per-page markers) and read that; only OCR genuinely weak pages. HW sheets often cite only textbook problem numbers; the local textbook PDF yields exact statements + printed answers via full-text cache + tag grep. Both recipes: references/lecture-slide-extraction.md.
Pitfalls
- Assignment sheets override assumptions: HW PDF headers carry per-assignment authority (submission platform, co-instructors) and can contradict the course README/syllabus. Flag the discrepancy to the user before advising where to submit; reconcile the README once confirmed.
- Textbook answers are for verification, not the solve: the prof's rubric (points off for wrong answer, no work shown, wrong/missing units) makes the work the grade. Never hand over textbook answers as a solution.
- Check answers carry units (engineering notation applies —
10 µAnot10). Unit slips are the cheapest points to lose on homework. - When the user flags a stale prerequisite ("years since I took calc"), embed a two-minute refresher at the point of use in a HW problem (e.g. area-under-curve, polynomial integration for Ch 1-2 circuits) instead of a standalone lesson.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

