Teaches Lean 4 programming and Mathlib proofs through beginner-friendly Japanese tutoring.
npx skills add https://github.com/kohdice/dotfiles --skill lean-tutor-kohdice
Please help me install this Agent Skill. Skill: lean-tutor Source: https://github.com/kohdice/dotfiles/tree/main/config/agents/skills/lean-tutor Command: npx skills add https://github.com/kohdice/dotfiles --skill lean-tutor-kohdice
rw vs simp, Prop vs Bool, and term-mode vs tactic-mode proofs. - Toolchain and project guidance: Covers elan, lean-toolchain, Lake projects, and Mathlib setup including lake exe cache get, with version handling based on the latest stable Lean 4 release. - Use Case: A beginner asks why rw [Nat.zero_add] fails on the goal n + 0 = n; the tutor shows the goal state, explains that rw matches syntactic patterns, provides the fix with Nat.add_zero, and teaches how to avoid the mistake using exact?. ## Quick Start Ask the tutor to explain a Lean 4 concept or proof in beginner-friendly Japanese, for example by requesting an explanation of why a specific tactic proof works.High-intent search queries and answers about installing and using this skill.
Start with core Lean for functional programming concepts like List.map before adding Mathlib for mathematics. Use #check and #eval to explore types, read the infoview goal state after every tactic, and practice with small single-file examples that run with lean file.lean.
The rw tactic matches patterns syntactically, so Nat.zero_add (0 + n) cannot rewrite a goal containing n + 0 even though they are mathematically equal. Check that the lemma's left-hand side appears literally in the goal, or use exact? and rw? to find matching lemmas.
Core Lean suffices for functional programming with Nat, Int, List, and basic logic. Mathlib is required for real numbers, sets, groups, and tactics like norm_num, ring, and linarith, and it pins its own lean-toolchain version.
Term-mode proofs build the proof value directly, like ⟨hq, hp⟩ for a conjunction, while tactic mode with by interactively transforms the goal state step by step. Both produce the same checked proof term; tactic mode is usually easier for beginners to follow.
sorry closes any goal without a proof, so a theorem containing it proves nothing and Lean warns with 'declaration uses sorry'. It is acceptable as a placeholder while learning, but you should verify no sorry remains before trusting a result.
Install elan, run lake new my_project math to generate a project with Mathlib in lakefile.toml, then run lake exe cache get to download prebuilt Mathlib before lake build. The project's lean-toolchain must match Mathlib's pinned toolchain for the cache to work.
Arthur742RamosBuild, test, and debug Lean 4 projects using Lake.
trailofbitsWrites and reviews structured Lean 4 proofs following Mathlib conventions.
r-irbeRedirect Mathlib PR review standards to updated reference guides.
rainoftimeProve relational parametricity and derive free theorems for polymorphic functions.