[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-83323":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":9,"languages":9,"totalLinesOfCode":9,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":13,"subscribersCount":13,"size":13,"stars1d":12,"stars7d":14,"stars30d":14,"stars90d":13,"forks30d":13,"starsTrendScore":11,"compositeScore":15,"rankGlobal":9,"rankLanguage":9,"license":16,"archived":17,"fork":17,"defaultBranch":18,"hasWiki":19,"hasPages":17,"topics":20,"createdAt":9,"pushedAt":9,"updatedAt":21,"readmeContent":22,"aiSummary":9,"trendingCount":13,"starSnapshotCount":13,"syncStatus":23,"lastSyncTime":24,"discoverSource":25},83323,"claude-tutors","temorozov\u002Fclaude-tutors","temorozov","A collection of Claude skills for self-learners. Your AI becomes more than a tutor — it builds genuine understanding and teaches you how to think.",null,113,5,1,0,12,2.33,"MIT License",false,"main",true,[],"2026-06-12 02:04:33","# claude-tutors\n\nSkills that make Claude tutor the way the best human teachers do — guiding you to the answer instead of handing it over, so you actually learn it.\n\n## The problem they solve\n\nAI tutors have one failure mode: they explain *too* well. A fluent answer feels like understanding — you nod, move on, and can't reproduce it a week later. Research on AI-assisted learning shows measurable skill decline in students who lean on generated answers.\n\nThese skills are engineered against exactly that. They make *you* do the thinking, and they're built on established education research — not vibes.\n\n## The skills\n\n| Skill | What it is |\n|-------|-----------|\n| [`tutor`](.\u002Ftutor\u002F) | Universal tutor for any subject. The shared engine below, plus a *path mode* for \"I want to learn X.\" Used as the fallback when no specialist is installed. |\n| [`math-tutor`](.\u002Fmath-tutor\u002F) | Math, algebra through calculus, probability, and linear algebra — with a catalog of the specific misconceptions learners hold at each topic. |\n| [`programming-tutor`](.\u002Fprogramming-tutor\u002F) | Programming, language-agnostic. Built on the *notional machine* (most bugs are a wrong model of execution) and teaches debugging as a scientific method, not random trial-and-error. |\n\n## How they actually work\n\n- **A hint ladder, not an answer.** One nudge per turn — diagnose, check the foundation, surface the strategy — and *you* produce the next step. Producing it is what builds the skill; receiving it doesn't.\n- **It reads the situation first.** Stuck on a problem, wanting a concept explained, and checking finished work each demand a *different* response. Giving the wrong one is the most common way tutoring fails.\n- **No \"does that make sense?\"** — a yes\u002Fno question that everyone answers \"yes.\" Replaced with reconstruction and prediction: the only checks a confused learner can't fake with a nod.\n- **A misconception catalog per subject.** Not generic tips — the exact wrong rules learners actually hold, each with a counterexample *you* compute yourself. Believe `(a+b)² = a²+b²`? You work out `(1+1)² = 4 ≠ 2` and watch your own rule break. That sticks; being told you're wrong doesn't.\n\nGrounded in Skemp's relational understanding, du Boulay's notional machine, PRIMM, and cognitive load theory — sources in each skill's `references\u002F`.\n\n## Install\n\n1. Download a skill folder as a ZIP — open it (e.g. [`math-tutor`](.\u002Fmath-tutor\u002F)), click the green **Code** button → **Download ZIP**.\n2. In [claude.ai](https:\u002F\u002Fclaude.ai): **Settings → Customize → Skills → +** → **Upload a skill**, then upload the ZIP.\n3. Done — Claude uses it automatically when you study.\n\n## Contributing\n\nHave an idea for a skill? Open an issue or a pull request.\n",2,"2026-06-11 04:10:57","CREATED_QUERY"]