genpark-evol-instruct-prompt-complexity-enhancer-skill

An algorithmic prompt mutation engine implementing the WizardLM Evol-Instruct paradigm. Synthesizes high-complexity training data from simple seed instructions through structured mutation templates.

Architecture

flowchart TD
    Seed[Seed Instruction] --> Mutator[EvolInstructMutator]
    Mutator --> T1[Deepen Reasoning]
    Mutator --> T2[Add Constraints]
    Mutator --> T3[Concretize]
    Mutator --> T4[Increase Reasoning Steps]
    Mutator --> T5[In-Breadth Evolution]
    T2 --> Evaluator[Complexity Scoring Engine]
    Evaluator --> Mutated[High-Complexity Benchmark Dataset]

Features

• Deterministic Complexity Scoring: Quantifies lexical length, logical connectives, and constraints.
• 5 Mutation Vectors: Deepens reasoning, adds technical constraints, concretizes abstract ideas, expands step count, and expands cross-domain analogies.
• 100% Python Standard Library: Zero pip dependencies.
• Model Context Protocol (MCP): Native stdio server support.