learn-from-session
Skill0 starsChanged 21 days agoMITView on GitHub
Use when the user wants to learn or remember what was done in this coding session ("teach me what we just did", "turn this session into a course", "I don't want to forget how this works", "quiz me on what you changed"). Extracts the general concepts the session relied on (technologies, patterns, decisions, bugs and their fixes), proposes a short course, saves it to BrainyGym after the user approves it, and starts a first quiz. Never sends code, secrets or project-internal names.
5 other projects use a version of this.
Adopted by
Most recently taken up by
- notacp/learn-from-sessiontheir own version, first seen Aug 1, 2026
- Arcadi4/skillstheir own version, first seen Jul 6, 2026
- Erlemar/cayley-puzzlestheir own version, first seen Jul 5, 2026
- Erlemar/dswoktheir own version, first seen Jun 22, 2026
- doodledood/manifest-devtheir own version, first seen Jun 10, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- doodledood/manifest-dev★ 78same name
- Erlemar/dswok★ 33same name
- notacp/learn-from-session★ 17same name
- Arcadi4/skills★ 4same name
- Erlemar/cayley-puzzles★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository2 other files it gives its agents.
- brainygym-course-designSkill
- learn-codebaseSkill
The file
We split this skill into the rules it gives an agent
7 rules. 0 are common, 4 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
4
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 26
Look back over the session and list what the user would need to understand to do this work again without you.
- Distinctive: line 36
Never include source code, file contents, diffs, logs, stack traces, credentials, tokens, keys, URLs of internal systems, customer data or …
- Very specific: line 38
Never include project-internal names: repository, company, service, class, function, table or file names.
- Distinctive: line 42
If the session was about something sensitive (security vulnerabilities, unreleased products), say so and ask before going on.
- Distinctive: line 53
Ask: "Shall I save this to BrainyGym?" Adjust it with their feedback.
- Very specific: line 57
Call get_course_authoring_guide and follow it exactly (module size, reading, questions, the format save_course accepts).
- Very specific: line 66
If the user agrees, call start_quiz with the planId.
“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Learn from this coding session
Coding agents ship code faster than people absorb it. This skill turns what happened in the current session into a short BrainyGym course, so the user keeps the knowledge: BrainyGym schedules reviews of each concept with spaced repetition and puts them in the user's daily session on the web and the app.
Reply in the user's language. BrainyGym supports Spanish (es) and English (en); for any other
language, ask which of the two they prefer.
0. Check the connection
This skill needs the BrainyGym tools (save_course, start_quiz, get_course_authoring_guide). If
they are not available, tell the user how to connect, then stop:
- Claude Code:
claude mcp add --transport http brainygym https://brainygym.com/mcp, then/mcpto sign in. - Other agents: add the remote MCP server
https://brainygym.com/mcp(Streamable HTTP) in the agent's MCP settings and sign in with a BrainyGym account (free to start, 18+).
1. Extract concepts, not code
Look back over the session and list what the user would need to understand to do this work again without you. Good concepts are general and reusable:
- the technologies, libraries and APIs that were used, and why;
- patterns and decisions (and the alternative that was rejected);
- the bugs that came up: the symptom, the cause and how it was diagnosed;
- commands and tools that mattered (in general form).
Privacy rules (mandatory):
- Never include source code, file contents, diffs, logs, stack traces, credentials, tokens, keys, URLs of internal systems, customer data or personal data.
- Never include project-internal names: repository, company, service, class, function, table or file
names. Rewrite every concept in general terms ("a Next.js route handler that streams a response", not
"
app/api/foo/route.tsin Acme's billing service"). - Short illustrative snippets are fine only if you write them from scratch as generic examples.
- If the session was about something sensitive (security vulnerabilities, unreleased products), say so and ask before going on.
2. Propose the course and wait for approval
Show the user a draft syllabus first and do not save anything yet:
- a course title and a one-line goal ("Understand X so I can do Y on my own");
- 2–5 modules in learning order, each with 2–5 concepts and one sentence on what it covers;
- a note that nothing from the code itself will be stored.
Ask: "Shall I save this to BrainyGym?" Adjust it with their feedback. Save only after a clear yes.
3. Write and save it
Call get_course_authoring_guide and follow it exactly (module size, reading, questions, the format
save_course accepts). Keep the readings general, apply the privacy rules again to every reading,
question and explanation, and save with save_course (a new UUID as requestId).
Then tell the user what was saved, share the course link from the tool result, and offer to start right away.
4. First practice
If the user agrees, call start_quiz with the planId. If the BrainyGym quiz view is shown, the user
answers there and you do not repeat the questions; otherwise ask them one at a time and grade each with
submit_quiz_answer. Never reveal the answer before the user answers and never invent grades.
Finish by mentioning that the reviews will come back in their BrainyGym daily session.
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.

