numina-lean-agent
A toolkit for Lean 4 theorem proving, offering lemma search, verification, and code repair.
Install
mkdir -p .claude/skills/numina-lean-agent && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/13143" && unzip -o skill.zip -d .claude/skills/numina-lean-agent && rm skill.zipInstalls to .claude/skills/numina-lean-agent
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Lean 4 theorem proving toolkit: search lemmas, verify proofs, repair/simplify code, and get LLM-assisted informal proofsKey capabilities
- →Search mathematical lemmas
- →Verify proofs using lean-check, verify-proof, disprove
- →Repair Lean proofs
- →Simplify theorems
- →Generate LLM-assisted informal proofs
- →Extract theorems from code
How it works
The skill provides an index of tools for searching, verifying, and transforming Lean 4 code, as well as LLM-powered tools for informal proof generation and discussion.
Inputs & outputs
When to use numina-lean-agent
- →Search mathematical lemmas
- →Verify proofs
- →Repair Lean proofs
- →Simplify theorems
About this skill
Numina Lean Agent — Skills Index
Skills
| Skill | Description |
|---|---|
| search | Search tools: leanexplore, loogle, leanfinder, leansearch, state-search, hammer-premise |
| verification | Verification: lean-check, verify-proof, disprove |
| code-transform | Code transforms: repair-proofs, simplify-theorems, sorry2lemma, extract-theorems |
| llm | LLM tools: informal_prover, discussion_partner, code_golf |
Environment variables
GEMINI_API_KEY— informal_prover (gemini generation, gemini verifier, gemini refinement), code_golf, discussion_partner (gemini)OPENAI_API_KEY— informal_prover (gpt generation, gpt verifier), discussion_partner (gpt)ANTHROPIC_API_KEY— informal_prover (claude verifier)AXLE_API_KEY— axle commands (verify-proof, disprove, sorry2lemma, etc.)
When not to use it
- →When the user is not working with Lean 4 theorem proving
- →When the user does not need code transformations for proofs
- →When the user does not need LLM-assisted proof generation
Prerequisites
Limitations
- →The skill is specifically for Lean 4 theorem proving.
- →It requires various API keys (GEMINI_API_KEY, OPENAI_API_KEY, ANTHROPIC_API_KEY, AXLE_API_KEY) for full functionality.
- →The skill's capabilities are limited to the listed tools (search, verification, code-transform, llm).
How it compares
This skill provides a specialized toolkit for Lean 4 theorem proving, integrating search, verification, code transformation, and LLM-assisted informal proof generation, unlike general programming tools.
Compared to similar skills
numina-lean-agent side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| numina-lean-agent (this skill) | 0 | 3mo | No flags | Advanced |
| unsloth | 15 | 8mo | No flags | Intermediate |
| llm-application-dev | 3 | 4mo | Review | Intermediate |
| book-sft-pipeline | 3 | 3mo | Review | Advanced |
Try saying
Example prompts that trigger this skill in your AI assistant.
You might also like
unsloth
zechenzhangAGI
Expert guidance for fast fine-tuning with Unsloth - 2-5x faster training, 50-80% less memory, LoRA/QLoRA optimization
llm-application-dev
skillcreatorai
Building applications with Large Language Models - prompt engineering, RAG patterns, and LLM integration. Use for AI-powered features, chatbots, or LLM-based automation.
book-sft-pipeline
muratcankoylan
This skill should be used when the user asks to "fine-tune on books", "create SFT dataset", "train style model", "extract ePub text", or mentions style transfer, LoRA training, book segmentation, or author voice replication.
fine-tuning-with-trl
davila7
Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works with HuggingFace Transformers.
huggingface-tokenizers
davila7
Fast tokenizers optimized for research and production. Rust-based implementation tokenizes 1GB in <20 seconds. Supports BPE, WordPiece, and Unigram algorithms. Train custom vocabularies, track alignments, handle padding/truncation. Integrates seamlessly with transformers. Use when you need high-performance tokenization or custom tokenizer training.
massgen-develops-massgen
massgen
Guide for using MassGen to develop and improve itself. This skill should be used when agents need to run MassGen experiments programmatically (using automation mode) OR analyze terminal UI/UX quality (using visual evaluation tools). These are mutually exclusive workflows for different improvement goals.