martamakes/42-exam-rank-42/estudio42

Estudio42 — 42 School Socratic Tutor

How to run it

From the project root, without changing directory first:

node skills/estudio42/cli.js

This is a thin CLI wrapper (cli.js) around the existing index.js module, which in turn uses detect.js (exercise detection), generator.js (tip generation from the subject file), and prompts.js (es/en strings). Running it:

• Reads or creates .estudio42/config.json (language, username) at the

project root — the nearest ancestor of the cwd that contains .git.

• Detects the exercise the user is currently working on from the cwd.
• Loads .estudio42/tips/rank-<NN>/level-<L>/<exercise>.md if it already

exists, or generates it from the exercise's subject file and saves it there.

It prints one self-contained block of instructions to stdout. Read it and follow it for the rest of the session — it already states whether an exercise was detected, the tutoring rules, the available session commands, and the tips/guide content for that exercise. Don't re-derive or paraphrase this behavior yourself; treat the printed block as authoritative for how to act, in whichever language (es/en) it comes back in.

Re-run it whenever the user says they've switched exercises ("siguiente ejercicio" / "next exercise"), since detection is cwd-based.

Adding a user tip

The script does not perform "add tip" edits itself — generator.js exports an addUserTip helper, but nothing currently calls it; treat it as unused. When the user says "añade mi tip: ..." / "add my tip: ...", make the edit yourself with your file-editing tools: open the tips file the script reported, find (or create) a ### @<username> (<YYYY-MM-DD>) heading under the ## 👤 Tips de Usuarios / ## 👤 User Tips section — username comes from .estudio42/config.json — and append the tip as a bullet under it. Never edit another user's subsection.

Known limitations

None currently tracked. subject.en.txt lookup and the es/en translation gaps (in both index.js's prompt assembly and generator.js's auto-generated concepts/complexity/resources) were fixed and verified end-to-end in both languages — if you spot Spanish leaking through while config.json's language is "en" (or vice versa), that's a regression, not a known gap; flag it.