Skip to content
Skills Library
Education · Learning & explanation

learn-session

/learn-session

Learn material the user does not understand ("learn X") — find the missing concepts out to the boundary of what they know, teach them, and write cards. Use when the user asks to learn or understand a paper, chapter, link or question.

Repository
loci ↗
Licence
MIT
Last repository change
10.11.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 10.11.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Learn material the user does not understand ("learn X") — find the missing concepts out to the boundary of what they know, teach them, and write cards. Use when the user asks to learn or understand a paper, chapter, link or question.

Triggers from the description

learn X
Original author description

Learn material the user does not understand ("learn X") — find the missing concepts out to the boundary of what they know, teach them, and write cards. Use when the user asks to learn or understand a paper, chapter, link or question.

How it works

  1. Recording
  2. Grading
  3. Steps
Original SKILL.md headings · 70 lines, 640 words
SKILL.md preview
---
name: learn-session
description: Learn material the user does not understand ("learn X") — find the missing concepts out to the boundary of what they know, teach them, and write cards. Use when the user asks to learn or understand a paper, chapter, link or question.
---

# Learn session

If you have not read them this session, read the framework rules (`loci rules`),
`RULES.md` and `STYLE.md` first.

The input is material the user does not understand: a paper, a chapter, a link or a
question. Build the knowledge needed to understand it, searching outward until every
path reaches something the user already knows.

## Recording

- Grades go to the review log with `loci learn log` (one JSON object on stdin):
  `{"card": "<card id>", "grade": "<grade>"}` for a card, or
  `{"concept": "<concept id>", "grade": "<grade>"}` for a question on a concept that
  has no card. A grade is `again`, `hard`, `good` or `easy`.
- The session's questions on the material, their answers and grades, are recorded in
  a dated journal entry for the session in `journal/learn/` (named as the framework
  rules say), not in the review log.
- `loci learn state <slug>` shows what is known about a concept.

## Grading

Before showing a question, fix its reference answer and a short rubric. Grade the
answer against the rubric: every criterion met is `good`, some met is `hard`, none
met is `again`. Show the grade before logging it. If the user disputes it, re-judge
from scratch without looking at the first grade, and log only the final grade.

## Steps

0. **Questions.** Draft questions covering the claims the material cannot be
   understood without; the user may edit, add or drop them. The user answers them
   before any teaching. Record questions, answers and grades in the session's journal
   entry.
1. **Identify.** File the material as a source (framework rules: `loci rules`,
   "Filing a source"). List the concepts it uses, with
   the passages that use them. Each is load-bearing (the material cannot be
   understood without it) or peripheral (it gets a one-line definition where it
   appears).
2. **Sort.** Pre-fill each load-bearing concept as know, shaky or don't know, from
   the knowledge base (slug, or a search with `rg`) and `loci learn state`. The user
   corrects each. Know and shaky are the boundary; don't-know concepts are expanded.
   For a concept already in the knowledge base, offer its `prereqs` as the next level.
3. **Expand.** For each don't-know concept, find external material that explains it
   (the introducing paper, a textbook chapter, documentation), file it as a source,
   list its load-bearing prerequisites and sort them as in step 2.
   Repeat until every branch reaches the boundary or the user stops. New concept
   notes record their `prereqs`. Watch for synonyms (duplicate expansion) and
   homonyms (a false "know").
4. **Plan and diagnose.** Write the plan as a view, ordered from the boundary
   inward, in the `views/` folder the vault keeps learning plans in (ask the user the
   first time, and record the answer in `RULES.md`). Create notes with `loci new` for
   every new or shaky concept that has none. Diagnose each shaky concept until it is
   settled: a clear pass makes it known, a clear fail makes it don't-know, a mixed
   result gets another item from a different angle. The user may settle any concept
   directly.
5. **Teach,** from the boundary inward, in plan order until the user ends the session
   (the plan keeps the place): a short direct explanation, a worked example only for
   unseen material, then a step the user does themselves. Never do that step for
   them. New notes link the fetched sources.
6. **Return to the material.** Link its passages to the new concepts. Ask the step-0
   questions again and record the new answers and grades in the journal entry.
7. **Cards.** Write cards in each concept's `## Cards` section until every idea worth
   remembering is covered, drawing on items the user answered; the user approves
   each. Split a proof into key-step and why cards. Write cards without ids, then run
   `loci block-ids <note> --section Cards`. The step-0 questions become cards on the
   closest concept.

Finish with `loci done`.

Versions

The database has no other skills named learn-session.

Database snapshot: 13.11.2026 · package version b3bfb2f8406c (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100tutor
    Use when the user wants to learn, study, or understand a topic — triggered by requests to teach, tutor, explain concepts, quiz, drill, or build understanding of any subject
  • 100learn
    Use when the user says "teach me", "help me understand", "I want to learn", "resume <topic>", "review notes for <topic>", or "why does this work". Also "teach as we go", "teach me as we go", "stop teaching as we go", "mark this", "teach me what I marked", or "teach me the…
  • 100teach
    Teach the user a new skill or concept, within this workspace. Use when the user wants to learn a concept or skill inside this workspace. Triggers on: teach me, help me learn, explain this properly.
  • 100teach-me
    Teach a topic, video, article, codebase change, spec, or transcript until the user demonstrably understands it, using staged instruction, a running checklist doc, and quizzes. Use whenever the user says "teach me", "explain this in detail", "help me understand X", "/teach-me",…
  • 100learn
    Entry point for learning anything. Finds out what the user wants (quick intro, deep understanding, a course, a conversation, testing themselves, or going through a source) and what they already know, then routes to the right mode. Use for "/learn X", "I want to learn X", "teach…
  • 100teach-me
    Use when the user asks to learn a concept or understand completed work through stepwise tutoring, checks for understanding, or a saved walkthrough or learning record.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.