mathematical-theorem-prover
Implement comprehensive mathematical theorem proving capabilities with SFT+GRPO training, MCP/A2A agent integration, and imatrix quantization protection to surpass Boreas-phi3.5-instinct-jp in formal proof generation and scientific discovery. Use when building mathematical reasoning systems, formal verification tools, or AI-assisted theorem proving environments.
npx skills add majiayu000/claude-skill-registry --skill mathematical-theorem-prover --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
# 数学的定理証明能力強化システム ## 概要 このスキルは、SO8T/AEGIS統合戦略に基づき、数学的定理証明能力の包括的強化を実装します。SFT+GRPO訓練戦略、MCP/A2A汎用AIエージェント統合、Imatrix量子化保護により、Boreas-phi3.5-instinct-jpを上回る形式的証明能力と科学的発見支援を実現します。 ## 核心戦略 ### 1. SFT+GRPO訓練戦略 #### Phase 1: Mathematical Foundation SFT ```python math_sft_config = { "base_model": "microsoft/wavecoder-ultra", # SO8Tベース "math_datasets": [ "Proof-Pile-2", # Llemmaスタイル数学コーパス "Lean-Workbook", # 形式証明データ "MATH", # 競技数学問題 "miniF2F" # 形式証明ベンチマーク ], "training_objective": "next_token_prediction + proof_verification", "math_weight": 0.7 # 数学データ重み付け } ``` #### Phase 2: GRPO Reinforcement for Theorem Proving ```python grpo_config = { "reward_functions": [ "formal_proof_correctness", # Lean/Isabelle検証 "proof_completeness", # サブゴール解決率 "mathematical_novelty", # 新規補題生成 "proof_efficiency" # 証明長最小化 ], "group_size": 8, # GRPOグループサイズ "theorem_proving_env": "Lean4-interactive", "max_proof_depth": 50, "synthetic_data_generation": True } ``` ### 2. MCP/A2A汎用AIエージェント統合 #### Mathematical Reasoning Agent Architecture ```python class MathematicalReasoningAgent: def __init__(self): self.theorem_prover = MCPTool("lean4-prover") self.symbolic_solver = MCPTool("sympy-solver") self.hypothesis_generator = MCPT
- 概要
- 核心戦略
- 1. SFT+GRPO訓練戦略
- 2. MCP/A2A汎用AIエージェント統合
- 3. Imatrix量子化保護データ戦略
- Boreas-phi3.5-instinct-jp上回り戦略
- 1. 性能比較分析
- 2. 具体的な上回り戦略
- 実装可能な学習データ構造化
- 1. データアーキテクチャ
- 2. データ生成パイプライン
- 3. 訓練戦略の実装
- 実行ワークフロー
- 1. データ収集フェーズ
python scripts/data/generate_mathematical_training_data.py \ python scripts/setup_formal_proving_environment.py \ python scripts/training/grpo_mathematical_training.py \ python scripts/agents/develop_mathematical_agents.py \ python scripts/quantization/apply_math_protected_quantization.py \
What does the mathematical-theorem-prover skill do?
Implement comprehensive mathematical theorem proving capabilities with SFT+GRPO training, MCP/A2A agent integration, and imatrix quantization protection to surpass Boreas-phi3.5-instinct-jp in formal proof generation and scientific discovery. Use when building mathematical reasoning systems, formal verification tools, or AI-assisted theorem proving environments.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill mathematical-theorem-prover --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.